BIM-based automatic splitting and drawing system for fabricated building components

Abstract

This invention discloses a BIM-based automatic component disassembly and drawing system for prefabricated buildings, comprising the following modules: a parameter extraction module, which extracts and binds physical constraints to obtain a parameterized dataset; a boundary optimization module, which adjusts region boundaries to obtain an optimal disassembly layout; a sub-component generation module, which disconnects cross-component connections and adds nodes to obtain independent sub-components; an intelligent drawing module, which introduces an improved quantum annealing algorithm to construct an anisotropic potential field Hamiltonian, and avoids collisions through global energy optimization to obtain two-dimensional drawings; a consistency verification module, which intercepts errors through virtual reverse assembly to obtain verification drawings; and a manufacturing delivery module, which packages and exports production data packages and QR codes to obtain a processing dataset. This invention improves the compliance of component disassembly and the clarity of drawing layout, achieving seamless delivery of design and manufacturing data.

Description

Technical Field

[0001] This invention relates to the field of digital design and intelligent manufacturing integration of prefabricated buildings, and in particular to a BIM-based automatic component disassembly and drawing system for prefabricated buildings. Background Technology

[0002] With the development of industrialized construction, BIM-based digital design of prefabricated buildings has become a core technology in the industry. Although existing technologies have basic model segmentation and projection drawing functions, they still have significant shortcomings in practical applications.

[0003] First, in the component disassembly stage, existing technologies lack a mechanism for systematically matching multiple physical constraints such as structural stress, formwork dimensions, transportation width and height limits, and weight thresholds. Adjusting the disassembly boundaries relies heavily on manual trial and error, making it difficult to achieve a multi-objective global optimum that minimizes component weight and maximizes formwork reuse rate. This often results in disassembled components exceeding limits or low formwork utilization.

[0004] Secondly, the arrangement of graphic elements in 2D drawing is a major pain point in existing technologies. Dimension lines, rebar annotations, and embedded part symbols are prone to overlapping and collisions. Traditional software uses simple heuristic algorithms or relies on manual dragging to avoid collisions, which is time-consuming and prone to omissions. More critically, existing algorithms lack directional constraints, and graphic element adjustments are prone to deviating from the projection area, disrupting the drawing logic, and failing to combine spatial features such as projection normals to guide collision avoidance calculations.

[0005] Furthermore, existing technologies suffer from a disconnect between node processing and verification delivery. Errors are prone to occur when replacing cross-boundary rebar with pipelines after disconnection; the lack of virtual reverse assembly verification methods often leads to problems such as alignment deviations between reserved rebar and sleeves, embedded junction boxes, and discrepancies between two-dimensional and three-dimensional dimensions being discovered only on-site; simultaneously, data barriers exist between design and manufacturing, making it difficult to export production data packages containing CNC codes and traceability QR codes with a single click, hindering the digitalization of the entire process.

[0006] Therefore, how to provide an automatic disassembly and drawing system for prefabricated building components based on BIM is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] One objective of this invention is to propose a BIM-based automatic component disassembly and drawing generation system for prefabricated buildings. This invention fully integrates key steps such as physical constraint parameter matching, multi-objective boundary optimization, automatic cross-boundary node supplementation, improved quantum annealing algorithm for collision avoidance and arrangement, and virtual reverse assembly verification. It constructs a component disassembly and drawing generation system with features including constraint binding standardization, dynamic adjustment of disassembly boundaries, intelligent generation of connection nodes, anisotropic potential field modulation of primitive arrangement, and full-process traceability of processing data. Figure 1This integrated process enables global intelligent optimization of component decomposition, drawing generation, and manufacturing delivery within a complex BIM model environment. By introducing an improved quantum annealing algorithm, this invention constructs a Hamiltonian for anisotropic potential field modulation based on the projection normal. It innovatively injects a directional potential field according to the projection normal vector, resulting in extremely low potential barriers along the projection normal direction and extremely high potential barriers in the perpendicular direction, thus limiting the tunneling direction. This achieves directional guided optimization for element collision avoidance and possesses advantages such as multi-dimensional matching of decomposition constraints, collision-free element arrangement, strong linkage between drawing and model, pre-emptive verification and error correction, and traceable delivery data. It significantly improves the automation level and drawing clarity of prefabricated construction drawings, effectively solving problems in existing technologies such as difficulty in considering multiple constraints during decomposition, difficulty in avoiding collisions due to element overlap, and disconnect between design and manufacturing data.

[0008] The BIM-based automatic disassembly and drawing system for prefabricated building components according to an embodiment of the present invention includes the following modules: The parameter extraction module is used to extract physical constraint parameters from the BIM model, match and bind geometric information with physical constraints, and obtain a standardized parameterized basic dataset. The boundary optimization module is used to select the splicing points and structural seams that are permissible under stress as the reference division points based on the normalized parameterized basic dataset, to perform preliminary region division, adjust the region boundaries, and obtain the optimal split boundary layout map. The sub-component generation module is used to identify connection relationships and disconnect cross-boundary steel bars and pipelines according to the optimal split boundary layout diagram, supplement and generate reserved anchorage ends or sleeve connection nodes, and at the same time assign the attribute information of the original component to the newly split sub-component and give it an independent identity identifier, thus obtaining the independently split sub-component. The intelligent drawing module is used to generate two-dimensional views based on independently split sub-components. It introduces an improved quantum annealing algorithm to construct the Hamiltonian of the anisotropic potential field modulation problem based on the projection normal. Through optimization calculation of global energy value minimization, it automatically detects and avoids overlapping collisions to obtain two-dimensional engineering drawings. The consistency verification module is used to reverse assemble all independent sub-components in virtual space according to their original positions and reserved hoisting gaps before outputting two-dimensional engineering drawings. If a collision or size discrepancy is found, the drawing will be intercepted and the error position will be indicated. The verification results will be obtained from the drawings and sub-components that have passed the verification. The manufacturing delivery module is used to package the verified drawings and the 3D models, 2D drawings and processing attributes of the sub-components, export the production data package, generate a tracking QR code, and obtain the processing dataset.

[0009] Optionally, the parameter extraction module specifically includes: By calling the data interface of the BIM model, the three-dimensional solid objects in the model are traversed, and the length, width and height values ​​of each solid object are read as geometric dimensions. At the same time, the material attribute text configured for each solid object is read as material information. Extract the preset structural stress limits, preset mold table dimensions, preset transportation width and height limits, and preset component weight thresholds, and aggregate them into physical constraint parameters; The length, width, and height values ​​in the read geometric dimensions are compared with the preset mold table size and preset transportation width and height limits in the physical constraint parameters. When the geometric dimension is less than or equal to the preset mold table size and less than or equal to the preset transportation width and height limits, it is determined that the dimensions match. Simultaneously, the estimated weight of the component is obtained by multiplying the material density parameter corresponding to the material information with the volume calculated from the geometric dimensions. The estimated weight of the component is then compared with the preset component weight threshold in the physical constraint parameters. When the estimated weight of the component is less than or equal to the preset component weight threshold, it is determined that the weight is matched. The allowable design stress corresponding to the material information is then compared with the preset structural stress limit in the physical constraint parameters. When the allowable design stress is less than or equal to the preset structural stress limit, it is determined that the stress is matched. The geometric dimensions and material information that are determined to be size matching, weight matching, and stress matching are merged with the corresponding physical constraint parameters such as structural stress limits, mold size, transportation width and height limits, and component weight thresholds and written into the same dataset. The geometric information and physical constraint conditions are matched and bound to obtain a standardized parameterized basic dataset.

[0010] Optionally, the boundary optimization module specifically includes: Based on the standardized parametric basic dataset, non-primary stress areas with design allowable stress less than or equal to the structural stress limit are found in the BIM model. Structural connection joints within the non-primary stress areas are extracted as structural joints, and splicing points and structural joints within the non-primary stress areas are marked as reference division points. Using the marked reference division points as cutting positions, the building space corresponding to the BIM model is divided along the horizontal and vertical directions, dividing the building space into multiple continuous preliminary area blocks, thus completing the preliminary area division of the building space. The weight of the region block is obtained by multiplying the volume calculated from the geometric dimensions of each preliminary region block by the material density parameter corresponding to the material information. The outer contour length and width parameters of each preliminary region block are extracted. The number of times the same outer contour length and width parameters appear is counted and divided by the total number of preliminary region blocks to obtain the template reuse rate. Calculate the weight of the current preliminary region block and the weight of the adjacent preliminary region block respectively. Subtract the weight of the adjacent preliminary region block from the weight of the current preliminary region block to obtain the weight difference. Determine whether the weight of the current preliminary region block is greater than the component weight threshold in the normalized parameterized basic dataset. If it is, move the corresponding reference division point towards the adjacent preliminary region block with a positive weight difference to reduce the geometric size of the current preliminary region block. Otherwise, keep the position of the current reference division point unchanged. Determine whether the template reuse rate of the current preliminary region block is lower than the preset utilization rate threshold. If so, modify the outer contour length and width parameters of the current preliminary region block to match the outer contour length and width parameters that appear most frequently in the statistics. Otherwise, keep the outer contour length and width parameters of the current preliminary region block unchanged. Repeatedly execute the steps of moving the reference division point and modifying the length and width parameters of the outer contour until the weight of all preliminary area blocks does not exceed the component weight threshold and the template reuse rate of all preliminary area blocks is not lower than the preset utilization rate threshold. Then, combine the adjusted area boundaries and reference division points to output the optimal split boundary layout diagram.

[0011] Optionally, the sub-component generation module specifically includes: Based on the baseline division points in the optimal split boundary layout diagram and the adjusted area boundary, a split section is generated in the BIM model along the adjusted area boundary. The three-dimensional solid objects intersecting with the split section are extracted as each solid component. The coordinates of the intersection points of the split section with the geometric center lines of the steel bars and pipelines inside the solid components are calculated. The steel bar attributes and pipeline attributes on both sides of the intersection point coordinates are read and recorded to obtain the connection relationship between the steel bars and pipelines across the components. The steel bars whose geometric center lines intersect with the cutting section are marked as cross-boundary steel bars, and the pipelines whose geometric center lines intersect with the cutting section are marked as cross-boundary pipelines. At the coordinates of the intersection point between the cross-boundary steel bars and the cross-boundary pipelines and the cutting section, the three-dimensional geometric models of the cross-boundary steel bars and the cross-boundary pipelines are broken, so that the cross-boundary steel bars and the cross-boundary pipelines are separated into two independent parts at the intersection point coordinates. At the endpoint where the cross-boundary reinforcement is broken, the diameter and material information of the cross-boundary reinforcement are read. When the diameter of the cross-boundary reinforcement is greater than or equal to the preset diameter threshold, a sleeve connection node with the same diameter as the cross-boundary reinforcement is generated on the outside of the endpoint. When the diameter of the cross-boundary reinforcement is less than the preset diameter threshold, an outwardly extending reinforcement segment is generated on the outside of the endpoint as a reserved anchoring end. At the endpoint where the cross-boundary pipeline is broken, the cross-sectional dimension of the cross-boundary pipeline is read, and a pipeline interface with the same cross-sectional dimension as the cross-boundary pipeline is generated on the outside of the endpoint as a pipeline connection node. Extract the geometric dimensions and material information recorded in the original solid component before disconnection. Copy the geometric dimensions and material information and write it into the attribute data table of each newly disconnected sub-component. Generate a unique alphanumeric string for each newly split sub-component according to the preset encoding rules. Write the string into the attribute data table of the sub-component as an independent identifier, thus obtaining the independently split sub-component.

[0012] Optionally, the intelligent drawing module specifically includes: The normal vectors of the outer contour surfaces are extracted from the three-dimensional geometric models of the independently decomposed sub-components. The area value of each outer contour surface is multiplied by the corresponding normal vector, and the sum is divided by the total area of ​​all outer contour surfaces to calculate the area-weighted average normal vector. The direction perpendicular to the area-weighted average normal vector is determined as the main force-bearing surface direction of the independent sub-component. The angle perpendicular to the main force-bearing surface direction and with the largest projected area is calculated as the optimal viewing angle. The independent sub-component is orthogonally projected along the optimal viewing angle to generate a two-dimensional view. On the 3D geometric model of the independently split sub-components, extract the coordinates of all vertices, the midpoint coordinates of each edge line, and the endpoint coordinates of the arc curves. Summarize these coordinates and calculate them as the coordinates of the key points of the 3D model. Extract the coordinates of the leader points of the dimension annotations on the drawings of the independently split sub-components in the 2D view. Establish a one-to-one mapping link relationship between the leader point coordinates and the corresponding 3D model key point coordinates and bind them. When the coordinates of the 3D model key points are modified, calculate the coordinate difference synchronously according to the link relationship and update the value of the dimension annotation. Extract the bounding rectangles of the dimension lines, rebar markings, and embedded part symbols on the drawing. Map each bounding rectangle to a discrete state variable. Define the x-coordinate and y-coordinate of the center point of each discrete state variable and the length and width parameters of the bounding rectangle as the position state of the qubit. Use the initial arrangement of the dimension lines, rebar markings, and embedded part symbols in the two-dimensional view as the initial position state of the qubit. An improved quantum annealing algorithm is introduced to calculate the projection direction vector of the optimal viewing angle of the two-dimensional view as the normal vector generated by the projection. The direction of the parallel component of each discrete state variable along the normal vector generated by the projection is defined as the tunneling allowable direction, and the direction of the component perpendicular to the normal vector generated by the projection is defined as the tunneling restrictive direction. The initial Hamiltonian is calculated, and the problem Hamiltonian is calculated using the local potential field strength and the interaction potential field strength. The anisotropic potential field modulation problem Hamiltonian based on the projection normal is constructed to obtain the total Hamiltonian of the system. The optimization calculation is performed in the continuous evolution step to output the final position coordinates of all discrete state variables. The dimension lines, rebar markings, and embedded part symbols are rearranged in the 2D view according to the final position coordinates. The system automatically detects and avoids overlapping and collisions between dimension lines, rebar markings, and embedded part symbols to produce 2D engineering drawings.

[0013] Optionally, the calculation of the initial Hamiltonian, and the calculation of the problem Hamiltonian using the local potential field strength and the interaction potential field strength, constructing the anisotropic potential field modulation problem Hamiltonian based on the projection normal, obtaining the total Hamiltonian of the system, and performing optimization calculations in the continuous evolution step to output the final position coordinates of all discrete state variables, specifically includes: The initial Hamiltonian is obtained by multiplying the transverse tunneling amplitude along the permitted tunneling direction with the preset transverse Pauli matrix operator of the corresponding discrete state variable, adding the transverse tunneling amplitude along the restricted tunneling direction with the preset longitudinal Pauli matrix operator of the corresponding discrete state variable, and summing the product results of all discrete state variables. The local potential field strength is obtained by multiplying the reciprocal of the shortest distance from the corresponding discrete state variable to the outer contour of the independent split sub-component with the preset attraction coefficient. The overlapping area is obtained by multiplying the overlap length of the bounding rectangles of every two discrete state variables on the horizontal and vertical axes. When the overlapping area is greater than zero, the interaction potential field strength is set to a preset positive penalty value. When the overlapping area is equal to zero, the interaction potential field strength is set to zero. The Hamiltonian of the problem is obtained by multiplying the local potential field strength of each discrete state variable with the corresponding preset Pauli Z operator and summing them, and by multiplying the interaction potential field strength between each pair of discrete state variables with the corresponding two preset Pauli Z operators and summing them. This constructs the Hamiltonian of the anisotropic potential field modulation problem based on the projection normal. Based on the normal vector generated by the projection, a directional potential field is injected. The lateral tunneling amplitude in the tunneling allowable direction is set to an extremely low barrier value, and the lateral tunneling amplitude in the tunneling restricted direction is set to an extremely high barrier value. This makes the barrier of the primitive along the projection normal direction extremely low and the barrier in the vertical direction extremely high, thus restricting the tunneling direction. The initial Hamiltonian is added to the problem Hamiltonian to obtain the total Hamiltonian of the system. An adiabatic evolution operation is performed on the total Hamiltonian of the system, and the transverse tunneling amplitude is gradually reduced to zero by multiplying the initial maximum value by a preset attenuation coefficient with the time step. At the same time, the local potential field strength and the interaction potential field strength are gradually increased to the target maximum value by multiplying the preset growth coefficient from zero. In each evolution step, the position and state of all current discrete state variables are combined into a column vector. The matrix corresponding to the total Hamiltonian of the system is multiplied by the column vector. The sum of squares of each element in the resulting column vector is calculated as the feature energy value of the current step. The feature energy value of the current step is subtracted from the feature energy value of the previous step, and the absolute value is taken as the energy difference. The energy difference is compared with a preset convergence threshold. When the energy difference is less than the preset convergence threshold, the evolution stops. The position state of all discrete state variables is read. The current feature energy value is compared with the lowest feature energy value in the history. If the current feature energy value is less than the lowest feature energy value in the history, the current position state is overwritten and saved as the final position coordinates. The final position coordinates of all discrete state variables are calculated and output through global energy minimization optimization.

[0014] Optionally, the consistency verification module specifically includes: Read the three-dimensional geometric model and corresponding independent identity of all independent sub-components in the virtual space, extract the original coordinate position of each independent sub-component recorded before disassembly as the original position, and inject a gap of preset width between the outer contours of all independent sub-components as a reserved hoisting gap. Each independent sub-component is moved and assembled in reverse in virtual space according to its original position and the reserved hoisting gap. The three-dimensional geometric model surfaces of adjacent independent sub-components after reverse movement and assembly are extracted, the intersection volume between adjacent surfaces is calculated, and a geometric collision is determined when the intersection volume is greater than zero. The position where the intersection volume is greater than zero is recorded as the geometric collision position. In the independent disassembled sub-components after reverse movement assembly, read the three-dimensional coordinates of the reserved anchorage end and the sleeve connection node generated at the break of the cross-boundary reinforcement, calculate the spatial distance between the center point of the reserved anchorage end and the center point of the corresponding sleeve connection node, and determine that the reserved reinforcement and the sleeve are not aligned when the spatial distance is greater than the preset alignment tolerance, and record the position where the spatial distance is greater than the preset alignment tolerance. Read the three-dimensional coordinates of the pipeline connection node and the corresponding pre-embedded junction box interface generated at the cross-border pipeline break point, calculate the spatial distance between the center point of the pipeline connection node and the center point of the pre-embedded junction box interface, and determine that the reserved pipeline and the pre-embedded junction box are not aligned when the spatial distance is greater than the preset alignment tolerance, and record the position where the spatial distance is greater than the preset alignment tolerance. Extract the values ​​of all dimensions in the 2D engineering drawings as 2D dimension data, read the coordinates of the corresponding key points in the 3D model and calculate the spatial distance between the two points as 3D dimension data, subtract the 2D dimension data and 3D dimension data at the same location and take the absolute value. When the absolute value is greater than the preset dimension error threshold, it is determined that the dimensions of the 2D engineering drawings and the dimensions of the 3D model are inconsistent, and the position of the dimension annotation corresponding to the absolute value being greater than the preset dimension error threshold is recorded. The system summarizes the locations of geometric collisions, misalignments between reserved reinforcing bars and sleeves, misalignments between reserved pipelines and embedded junction boxes, and dimensions inconsistencies between the dimensions on the 2D engineering drawings and the dimensions on the 3D model as error locations. It intercepts and outputs the 2D engineering drawings and prompts the error locations. If no error is found, it is determined that the verification is correct and the verification of the drawings and sub-components is passed.

[0015] Optionally, the manufacturing delivery module specifically includes: The verified drawings, the 3D model of the sub-component, the 2D drawings, and the processing attributes are merged into the same file and packaged. According to the preset data format requirements of the factory production system, the processing attributes are converted into a material cutting table, the steel bar outline in the 2D drawings is converted into steel bar processing data, and the processing trajectory of the 3D model is converted into CNC machine tool processing code. The material cutting list, rebar processing data, and CNC machine tool processing code are merged and exported as a production data package with one click. The unique identifier of each independent sub-component is read and converted into a QR code image. The QR code image is then attached to the corresponding production data package to obtain the processing dataset.

[0016] The beneficial effects of this invention are: This invention extracts the geometric dimensions and material information of the BIM model and combines it with structural stress limits, formwork dimensions, transportation width and height limits, and component weight thresholds for multi-constraint matching and binding. Addressing the lack of multi-dimensional physical constraints in prefabricated component disassembly, it employs boundary optimization to move the benchmark division point and adjust the outer contour parameters, achieving a dual-objective optimization of minimizing component weight and maximizing formwork reuse, generating an optimal disassembly boundary layout. By identifying cross-boundary reinforcement and pipelines along the disassembly boundary and performing a break operation, it automatically generates sleeve connection nodes or reserved anchorage ends based on diameter thresholds, completing intelligent supplementation and attribute information inheritance of connection nodes. During the intelligent drawing generation stage... By mapping drawing primitives to discrete state variables and introducing an improved quantum annealing algorithm, a Hamiltonian for anisotropic potential field modulation based on projection normals is constructed. Combined with numerical injection of directional potential fields at extremely low and high potential barriers to constrain tunneling directions, adiabatic evolution and global energy minimization optimization calculations are performed, outputting the final collision-free position coordinates of the primitives. In the consistency verification stage, reverse assembly is performed using reserved hoisting gaps. Geometric collisions, misalignment of embedded parts, and 2D / 3D dimensional deviations are accurately identified through intersection volume and coordinate distance calculations, intercepting erroneous drawings. Furthermore, through format conversion and QR code generation, one-click export of production data packages and full-process traceability are achieved. Ultimately, this system enables automatic multi-constraint splitting of prefabricated building components, directional collision-avoidance arrangement of drawing primitives, and closed-loop delivery of design and manufacturing data, effectively improving the compliance of component splitting, the intelligence of 2D drawing output, and the accuracy of model verification. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of the BIM-based automatic disassembly and drawing system for prefabricated building components proposed in this invention. Figure 2 This is a flowchart of the drawing primitive collision avoidance arrangement and consistency verification based on the improved quantum annealing algorithm proposed in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figures 1-2 The BIM-based automatic component breakdown and drawing generation system for prefabricated buildings includes the following modules: The parameter extraction module is used to extract geometric dimensions, material information, and physical constraint parameters such as structural stress limits, formwork dimensions, transportation width and height limits, and component weight thresholds from the BIM model. It matches and binds geometric information with physical constraints to obtain a standardized parameterized basic dataset. The boundary optimization module is used to select stress-allowed splicing points and structural joints in the BIM model as benchmark division points based on the standardized parametric basic dataset. This allows for the initial division of the building space into regions. The region boundaries are then repeatedly adjusted according to the principles of minimizing component weight and maximizing template reuse to obtain the optimal split boundary layout. The sub-component generation module is used to identify and record the connection relationships of each entity component in the BIM model and the steel bars and pipelines across components according to the optimal split boundary layout diagram. It disconnects the steel bars and pipelines across the boundary at the split boundary and automatically generates reserved anchorage ends or sleeve connection nodes. At the same time, it assigns the attribute information of the original component to the newly split sub-component and gives it an independent identity identifier, thus obtaining the independently split sub-component. The intelligent drawing module is used to calculate the optimal viewing angle based on the shape and main stress surface of the independently split sub-components and project to generate a two-dimensional view. It binds the dimension annotations on the drawing with the key points of the three-dimensional model and introduces an improved quantum annealing algorithm to construct the Hamiltonian of the anisotropic potential field modulation problem based on the projection normal. It maps the dimension lines, rebar annotations and embedded part symbols into discrete state variables and injects a directional potential field according to the normal vector generated by the projection. This makes the potential barrier of the primitives extremely low along the projection normal direction and extremely high in the vertical direction to limit their tunneling direction. Through global energy minimization optimization calculation, it automatically detects and avoids the overlapping collisions between the dimension lines, rebar annotations and embedded part symbols in the drawing, and produces a two-dimensional engineering drawing that is linked with the three-dimensional model and has a clear layout. The consistency verification module is used to reverse assemble all independent sub-components in virtual space according to their original positions and reserved hoisting gaps before outputting two-dimensional engineering drawings. It checks for geometric collisions and whether the reserved steel bars are aligned with the sleeves and embedded junction boxes. At the same time, it compares whether the dimensions of the two-dimensional engineering drawings are consistent with the dimensions of the three-dimensional model. If a collision or size discrepancy is found, the drawing is intercepted and the error position is indicated. The verification results show that the drawings and sub-components have passed the verification and are correct. The manufacturing delivery module is used to package the verified drawings and the 3D models, 2D drawings and processing attributes of the sub-components, and export the production data package containing the material cutting list, steel bar processing data and CNC machine tool processing code with one click, according to the data format requirements of the factory production system. It also generates a unique tracking QR code for each independently split sub-component, resulting in a processing dataset that can be directly used for factory production and manufacturing and can be traced throughout the entire process.

[0020] This invention significantly improves the compliance of prefabricated building component disassembly and the accuracy of intelligent drawing generation. Through parameter extraction and boundary optimization modules, it achieves deep integration of multi-dimensional physical constraints and geometric information, automatically finding the disassembly boundary that minimizes component weight and maximizes template utilization, effectively avoiding excessive disassembly. In the intelligent drawing generation stage, an improved quantum annealing algorithm is introduced to construct an anisotropic potential field modulated Hamiltonian. This directional potential field precisely restricts the tunneling direction of graphic elements, achieving global collision-free automatic arrangement of dimension lines and embedded part symbols, completely overcoming the technical barrier of avoiding overlapping dense graphic elements. The consistency verification module forms a rigorous closed-loop error correction mechanism through virtual reverse assembly and 2D / 3D dimension comparison, reducing assembly deviations and rework rates to extremely low levels. Finally, the manufacturing delivery module exports traceable production data packages with one click. This invention establishes a closed-loop data process from BIM design to factory manufacturing, demonstrating strong engineering practicality and significantly improving the intelligence level of complex component disassembly, drawing generation, and processing delivery.

[0021] In this embodiment, the parameter extraction module specifically includes: By calling the data interface of the BIM model, the three-dimensional solid objects in the model are traversed, and the length, width and height values ​​of each solid object are read as geometric dimensions. At the same time, the material attribute text configured for each solid object is read as material information. The preset structural stress limit, preset mold table size, preset transportation width and height limit, and preset component weight threshold are extracted and aggregated into physical constraint parameters. The preset structural stress limit is the maximum allowable stress that the component can withstand, which is 25 MPa. The preset mold table size is a standard steel mold table with a length of 12 meters and a width of 3.5 meters. The preset transportation width and height limit is the maximum allowable width of 3 meters and a height of 4.5 meters for road transportation. The preset component weight threshold is the maximum lifting capacity of the hoisting machinery, which is 15 tons. The length, width, and height values ​​in the read geometric dimensions are compared with the preset mold table size and preset transportation width and height limits in the physical constraint parameters. When the geometric dimension is less than or equal to the preset mold table size and less than or equal to the preset transportation width and height limits, it is determined that the dimensions match. Simultaneously, the estimated weight of the component is obtained by multiplying the material density parameter corresponding to the material information with the volume calculated from the geometric dimensions. The estimated weight of the component is then compared with the preset component weight threshold in the physical constraint parameters. When the estimated weight of the component is less than or equal to the preset component weight threshold, it is determined that the weight is matched. The allowable design stress corresponding to the material information is then compared with the preset structural stress limit in the physical constraint parameters. When the allowable design stress is less than or equal to the preset structural stress limit, it is determined that the stress is matched. The geometric dimensions and material information that are determined to be size matching, weight matching, and stress matching are merged with the corresponding physical constraint parameters such as structural stress limits, mold size, transportation width and height limits, and component weight thresholds and written into the same dataset. The geometric information and physical constraint conditions are matched and bound to obtain a standardized parameterized basic dataset.

[0022] In this embodiment, the boundary optimization module specifically includes: Based on the standardized parametric basic dataset, non-primary stress areas with design allowable stress less than or equal to the structural stress limit are found in the BIM model. Structural connection joints within the non-primary stress areas are extracted as structural joints, and splicing points and structural joints within the non-primary stress areas are marked as reference division points. Using the marked reference division points as cutting positions, the building space corresponding to the BIM model is divided along the horizontal and vertical directions, dividing the building space into multiple continuous preliminary area blocks, thus completing the preliminary area division of the building space. The weight of the region block is obtained by multiplying the volume calculated from the geometric dimensions of each preliminary region block by the material density parameter corresponding to the material information. The outer contour length and width parameters of each preliminary region block are extracted. The number of times the same outer contour length and width parameters appear is counted and divided by the total number of preliminary region blocks to obtain the template reuse rate. Calculate the weight of the current preliminary region block and the weight of the adjacent preliminary region block respectively. Subtract the weight of the adjacent preliminary region block from the weight of the current preliminary region block to obtain the weight difference. Determine whether the weight of the current preliminary region block is greater than the component weight threshold in the normalized parameterized basic dataset. If it is, move the corresponding reference division point towards the adjacent preliminary region block with a positive weight difference to reduce the geometric size of the current preliminary region block. Otherwise, keep the position of the current reference division point unchanged. Determine whether the template reuse rate of the current preliminary region block is lower than a preset utilization rate threshold. If so, modify the outer contour length and width parameters of the current preliminary region block to match the outer contour length and width parameters that appear most frequently in the statistics. Otherwise, keep the outer contour length and width parameters of the current preliminary region block unchanged. The preset utilization rate threshold is 60%. Repeatedly execute the steps of moving the reference division point and modifying the length and width parameters of the outer contour until the weight of all preliminary area blocks does not exceed the component weight threshold and the template reuse rate of all preliminary area blocks is not lower than the preset utilization rate threshold. Then, combine the adjusted area boundaries and reference division points to output the optimal split boundary layout diagram.

[0023] In this embodiment, the sub-component generation module specifically includes: Based on the baseline division points in the optimal split boundary layout diagram and the adjusted area boundary, a split section is generated in the BIM model along the adjusted area boundary. The three-dimensional solid objects intersecting with the split section are extracted as each solid component. The coordinates of the intersection points of the split section with the geometric center lines of the steel bars and pipelines inside the solid components are calculated. The steel bar attributes and pipeline attributes on both sides of the intersection point coordinates are read and recorded to obtain the connection relationship between the steel bars and pipelines across the components. The steel bars whose geometric center lines intersect with the cutting section are marked as cross-boundary steel bars, and the pipelines whose geometric center lines intersect with the cutting section are marked as cross-boundary pipelines. At the coordinates of the intersection point between the cross-boundary steel bars and the cross-boundary pipelines and the cutting section, the three-dimensional geometric models of the cross-boundary steel bars and the cross-boundary pipelines are broken, so that the cross-boundary steel bars and the cross-boundary pipelines are separated into two independent parts at the intersection point coordinates. At the endpoint where the cross-boundary reinforcement is broken, the diameter and material information of the cross-boundary reinforcement are read. When the diameter of the cross-boundary reinforcement is greater than or equal to a preset diameter threshold, a sleeve connection node with the same diameter as the cross-boundary reinforcement is generated on the outside of the endpoint. When the diameter of the cross-boundary reinforcement is less than the preset diameter threshold, an outwardly extending reinforcement segment is generated on the outside of the endpoint as a reserved anchoring end. At the endpoint where the cross-boundary pipeline is broken, the cross-sectional dimension of the cross-boundary pipeline is read. A pipeline interface with the same cross-sectional dimension as the cross-boundary pipeline is generated on the outside of the endpoint as a pipeline connection node. The preset diameter threshold is 14 mm. Extract the geometric dimensions and material information of the original solid component before disconnection, copy the geometric dimensions and material information and write it into the attribute data table of each newly disconnected sub-component. Generate a unique string of letters and numbers for each newly split sub-component according to the preset encoding rules, and write the string into the attribute data table of the sub-component as an independent identifier to obtain the independently split sub-component. The preset encoding rules are: project number plus hyphen plus component type abbreviation plus four incrementing numbers.

[0024] In this embodiment, the intelligent drawing module specifically includes: The normal vectors of the outer contour surfaces are extracted from the three-dimensional geometric models of the independently decomposed sub-components. The area value of each outer contour surface is multiplied by the corresponding normal vector, and the sum is divided by the total area of ​​all outer contour surfaces to calculate the area-weighted average normal vector. The direction perpendicular to the area-weighted average normal vector is determined as the main force-bearing surface direction of the independent sub-component. The angle perpendicular to the main force-bearing surface direction and with the largest projected area is calculated as the optimal viewing angle. The independent sub-component is orthogonally projected along the optimal viewing angle to generate a two-dimensional view. On the 3D geometric model of the independently split sub-components, extract the coordinates of all vertices, the midpoint coordinates of each edge line, and the endpoint coordinates of the arc curves. Summarize these coordinates and calculate them as the coordinates of the key points of the 3D model. Extract the coordinates of the leader points of the dimension annotations on the drawings of the independently split sub-components in the 2D view. Establish a one-to-one mapping link relationship between the leader point coordinates and the corresponding 3D model key point coordinates and bind them. When the coordinates of the 3D model key points are modified, calculate the coordinate difference synchronously according to the link relationship and update the value of the dimension annotation. Extract the bounding rectangles of the dimension lines, rebar markings, and embedded part symbols on the drawing. Map each bounding rectangle to a discrete state variable. Define the x-coordinate and y-coordinate of the center point of each discrete state variable and the length and width parameters of the bounding rectangle as the position state of the qubit. Use the initial arrangement of the dimension lines, rebar markings, and embedded part symbols in the two-dimensional view as the initial position state of the qubit. An improved quantum annealing algorithm is introduced to calculate the projection direction vector of the optimal viewing angle of the two-dimensional view as the normal vector generated by the projection. The direction of the parallel component of each discrete state variable along the normal vector generated by the projection is defined as the tunneling allowable direction, and the direction of the component perpendicular to the normal vector generated by the projection is defined as the tunneling restrictive direction. The initial Hamiltonian is calculated, and the problem Hamiltonian is calculated using the local potential field strength and the interaction potential field strength. The anisotropic potential field modulation problem Hamiltonian based on the projection normal is constructed to obtain the total Hamiltonian of the system. The optimization calculation is performed in the continuous evolution step to output the final position coordinates of all discrete state variables. The dimension lines, rebar labels, and embedded part symbols are rearranged in the 2D view according to the final position coordinates. The system automatically detects and avoids overlapping and collisions between dimension lines, rebar labels, and embedded part symbols, resulting in 2D engineering drawings that are linked to the 3D model and have a clear layout.

[0025] In this embodiment, the initial Hamiltonian is calculated, and the problem Hamiltonian is calculated using the local potential field strength and the interaction potential field strength. An anisotropic potential field modulation problem Hamiltonian based on the projection normal is constructed to obtain the total system Hamiltonian. A continuous evolution step is then performed to optimize and output the final position coordinates of all discrete state variables. Specifically, this includes: The initial Hamiltonian is obtained by multiplying the transverse tunneling amplitude along the permitted tunneling direction with the preset transverse Pauli matrix operator of the corresponding discrete state variable, and then multiplying the transverse tunneling amplitude along the restricted tunneling direction with the preset longitudinal Pauli matrix operator of the corresponding discrete state variable. The product results of all discrete state variables are then summed. The preset transverse Pauli matrix operator is a preset two-row, two-column matrix with the first row and first column having a value of zero, the first row and second column having a value of one, the second row and first column having a value of one, and the second row and second column having a value of zero. The preset longitudinal Pauli matrix operator is a preset two-row, two-column matrix with the first row and first column having a value of zero, the first row and second column having a value of negative one, the second row and first column having a value of one, and the second row and second column having a value of zero. The local potential field strength is obtained by multiplying the reciprocal of the shortest distance from the corresponding discrete state variable to the outer contour of the independent split sub-component with a preset attraction coefficient. The overlapping area is obtained by multiplying the overlap length of the bounding rectangles of every two discrete state variables on the horizontal and vertical axes. When the overlapping area is greater than zero, the interaction potential field strength is set to a preset positive penalty value. When the overlapping area is equal to zero, the interaction potential field strength is set to zero. The preset attraction coefficient is 0.5 and the preset positive penalty value is 5.0. The Hamiltonian of the problem is obtained by multiplying the local potential field strength of each discrete state variable with the corresponding preset Pauli Z operator and summing the results, and by multiplying the interaction potential field strength between each pair of discrete state variables with the corresponding two preset Pauli Z operators and summing the results. This constructs the Hamiltonian of the anisotropic potential field modulation problem based on the projection normal. The preset Pauli Z operator is a preset two-row two-column matrix with the first row and first column having a value of 1, the first row and second column having a value of 0, the second row and first column having a value of 0, and the second row and second column having a value of -1. Based on the normal vector generated by projection, a directional potential field is injected. The lateral tunneling amplitude in the tunneling allowable direction is set to an extremely low barrier value, and the lateral tunneling amplitude in the tunneling restricted direction is set to an extremely high barrier value. This makes the barrier of the primitive along the projection normal direction extremely low and the barrier in the vertical direction extremely high, thus restricting the tunneling direction. The extremely low barrier value is 0.1, and the extremely high barrier value is 10.0. The initial Hamiltonian is added to the problem Hamiltonian to obtain the total Hamiltonian of the system. An adiabatic evolution operation is performed on the total Hamiltonian of the system, so that the transverse tunneling amplitude gradually decreases to zero by multiplying the initial maximum value by a preset attenuation coefficient with the time step, while the local potential field strength and the interaction potential field strength are gradually increased to the target maximum value by multiplying the local potential field strength by a preset growth coefficient. The preset attenuation coefficient is 0.95 and the preset growth coefficient is 1.05. In each evolution step, the position and state of all current discrete state variables are combined into a column vector. The matrix corresponding to the total Hamiltonian of the system is multiplied by the column vector. The sum of squares of each element in the resulting column vector is calculated as the feature energy value of the current step. The feature energy value of the current step is subtracted from the feature energy value of the previous step, and the absolute value is taken as the energy difference. The energy difference is compared with a preset convergence threshold. When the energy difference is less than the preset convergence threshold, the evolution stops. The position state of all discrete state variables is read. The current feature energy value is compared with the lowest feature energy value in the history. If the current feature energy value is less than the lowest feature energy value in the history, the current position state is overwritten and saved as the final position coordinates. The final position coordinates of all discrete state variables are calculated and output through global energy minimization optimization. The preset convergence threshold is 0.01.

[0026] This invention achieves global collision-free arrangement and automatic obstacle avoidance optimization of graphic elements in complex drawings by introducing an improved quantum annealing algorithm combined with an anisotropic potential field modulation mechanism. Dimension lines, rebar markings, and embedded part symbols are mapped as discrete state variables. The area-weighted average normal vector is calculated to determine the optimal viewing angle, and the allowed and restricted tunneling directions are defined accordingly. A problem Hamiltonian containing local attractive forces and overlap penalty potential fields is constructed. During adiabatic evolution, extremely low potential barriers are injected along the projection normal direction, and extremely high potential barriers are injected perpendicularly, precisely restricting the tunneling direction of graphic elements. The final position coordinates are output through global energy minimization optimization. This invention can effectively avoid overlapping collisions of graphic elements under complex drawing conditions such as dense markings and multiple embedded parts intersections, completely eliminating manual movement and adjustment, and significantly improving the generation efficiency and layout clarity of 2D engineering drawings.

[0027] The improved quantum annealing algorithm of this invention is similar to the original quantum annealing algorithm in that both retain the core architecture of quantum annealing, namely, by constructing an initial Hamiltonian containing a transverse tunneling term and a problem Hamiltonian containing a potential field term, and according to the principle of adiabatic evolution, the transverse tunneling amplitude is gradually decayed to zero over time, while the problem potential field strength is gradually increased from zero to the target maximum value. Through the continuous evolution of the total Hamiltonian of the system, the state with the lowest global energy is found to solve the combinatorial optimization problem.

[0028] The difference lies in that this invention breaks the limitation of the original quantum annealing algorithm, which requires isotropic tunneling amplitudes for all qubits, and introduces an anisotropic potential field modulation mechanism based on the projection normal. While the original algorithm directly sets uniform tunneling parameters, this invention adds a projection normal vector analysis step to identify the allowed and restricted tunneling directions in the primitive arrangement. Based on this, a directional potential field is injected, setting the transverse tunneling amplitude of the allowed tunneling direction to an extremely low barrier value of 0.1, and the transverse tunneling amplitude of the restricted tunneling direction to an extremely high barrier value of 10.0. In the subsequent Hamiltonian calculation, different Pauli matrix operators are multiplied and accumulated with the corresponding tunneling amplitudes, forcing primitives to easily undergo quantum tunneling along the projection normal direction, while tunneling in the vertical direction is strictly restricted due to the extremely high barrier.

[0029] Based on the above improvements, the beneficial effects of this invention are that, through the modulation of directional potential field and differentiated potential barrier, the improved algorithm can adaptively guide the evolution path of quantum state to meet the physical requirements of sliding and avoiding elements along the projection direction and aligning them in the vertical direction in the layout of two-dimensional engineering drawings. This breaks the limitation of the original algorithm's isotropic tunneling causing disordered movement of elements, and realizes the constraint and orientation of the optimization search space. This design significantly reduces the search probability of invalid layout states, accelerates the convergence process of the algorithm to a legal solution, and can more accurately avoid overlapping collisions between dimension lines and embedded parts symbols, effectively improving the efficiency of automatic collision avoidance layout of drawing elements and the clarity of the drawing.

[0030] In this embodiment, the consistency verification module specifically includes: In the virtual space, read the three-dimensional geometric model and corresponding independent identity of all independent sub-components, extract the original coordinate position of each independent sub-component recorded before disassembly as the original position, and inject a gap of a preset width between the outer contours of all independent sub-components as a reserved hoisting gap, wherein the preset width gap is 50 mm. Each independent sub-component is moved and assembled in reverse in virtual space according to its original position and the reserved hoisting gap. The three-dimensional geometric model surfaces of adjacent independent sub-components after reverse movement and assembly are extracted, the intersection volume between adjacent surfaces is calculated, and a geometric collision is determined when the intersection volume is greater than zero. The position where the intersection volume is greater than zero is recorded as the geometric collision position. In the independently disassembled sub-components after reverse movement and assembly, the three-dimensional coordinates of the reserved anchorage end and the sleeve connection node generated at the break of the cross-boundary steel bar are read, the spatial distance between the center point of the reserved anchorage end and the center point of the corresponding sleeve connection node is calculated, and when the spatial distance is greater than the preset alignment tolerance, it is determined that the reserved steel bar and the sleeve are not aligned, and the position where the spatial distance is greater than the preset alignment tolerance is recorded. The preset alignment tolerance is 5 mm. Read the three-dimensional coordinates of the pipeline connection node and the corresponding pre-embedded junction box interface generated at the cross-border pipeline break point, calculate the spatial distance between the center point of the pipeline connection node and the center point of the pre-embedded junction box interface, and determine that the reserved pipeline and the pre-embedded junction box are not aligned when the spatial distance is greater than the preset alignment tolerance, and record the position where the spatial distance is greater than the preset alignment tolerance, which is 5 mm; Extract the values ​​of all dimensions in the two-dimensional engineering drawings as two-dimensional dimension data, read the coordinates of the corresponding key points in the three-dimensional model and calculate the spatial distance between the two points as three-dimensional dimension data, subtract the two-dimensional dimension data and the three-dimensional dimension data at the same location and take the absolute value. When the absolute value is greater than the preset dimension error threshold, it is determined that the dimensions of the two-dimensional engineering drawings and the dimensions of the three-dimensional model are inconsistent, and the position of the dimension annotation corresponding to the absolute value being greater than the preset dimension error threshold is recorded. The preset dimension error threshold is 2 mm. The system summarizes the locations of geometric collisions, misalignments between reserved steel bars and sleeves, misalignments between reserved pipelines and embedded junction boxes, and the locations of dimension annotations where the dimensions of the 2D engineering drawings and the 3D model are inconsistent as error locations. It intercepts and outputs the 2D engineering drawings and prompts the error locations. When no error is found, it is determined that the verification is correct and the verification drawings and sub-components are deemed to have passed the verification.

[0031] In this embodiment, the manufacturing and delivery module specifically includes: The verified drawings, the 3D model of the sub-component, the 2D drawings, and the processing attributes are merged and packaged into the same file. According to the preset data format requirements of the factory production system, the processing attributes are converted into a material cutting table, the steel bar outline in the 2D drawings is converted into steel bar processing data, and the processing trajectory of the 3D model is converted into CNC machine tool processing code. The preset data format requirements are in XML format. The material cutting list, rebar processing data, and CNC machine tool processing code are merged and exported as a production data package with one click. The unique identifier of each independent sub-component is read and converted into a QR code image. The QR code image is then attached to the corresponding production data package to obtain a processing dataset that can be directly used for factory production and is traceable throughout the entire process.

[0032] Example 1: To verify the feasibility of this invention in the field of automatic disassembly and drawing of prefabricated building components, this invention was applied to the intelligent construction platform of the brewing workshop and fermentation room of a raw wine production and storage expansion project in a certain city of a province. The project has a total construction area of ​​908,000 square meters, and a single project requires the installation of up to 260,000 fermentation tanks. The workshop requires a large number of prefabricated prefabricated tank protective covers, equipment foundation modules, track support components, and standardized fermentation partition wall panels. Under traditional construction methods, facing such a large volume of components, BIM disassembly relies heavily on manual experience, making it difficult to balance the dimensions of the formwork and the width and height limits for transportation, resulting in frequent rework due to component exceeding limits. At the same time, during the drawing stage, the dense embedded parts and steel reinforcement markings often cause overlapping and collision of elements. Manual collision avoidance is extremely time-consuming and prone to errors, seriously delaying the installation and commissioning progress of the fully mechanized intelligent transfer system (RGV and intelligent crane).

[0033] In this scenario, the method of this invention transforms the aforementioned multi-dimensional design and production constraints into a standardized parametric basic dataset. After the overall BIM brewing workshop model is imported, the system automatically extracts the geometric dimensions and material information of components such as the foundation module and track support, and strictly matches and binds them with the 25MPa structural stress limit, the 12m × 3.5m standard steel formwork table dimensions, the 3m wide × 4.5m high road transport width and height limit, and the 15-ton hoisting weight threshold. During the boundary optimization stage, the system automatically finds splicing points and structural joints in non-primary stress areas, dynamically and repeatedly adjusting the area boundaries based on the principle of minimizing component weight and ensuring a template reuse rate of no less than 60%, outputting the optimal split boundary layout diagram. Subsequently, the system automatically disconnects cross-boundary reinforcing bars and pipelines at the split boundaries, automatically generating sleeve connection nodes or reserved anchoring ends based on a 14mm diameter threshold, ensuring precise docking of subsequent RGV track splicing and equipment installation, and assigning unique identification to sub-components. In the crucial intelligent drawing generation stage, the system calculates the area-weighted average normal vector to determine the optimal viewing angle for orthogonal projection, and maps dimension lines, rebar annotations, and embedded part symbols as discrete state variables. The core lies in introducing an improved quantum annealing algorithm, constructing a Hamiltonian for anisotropic potential field modulation based on the projection normal, injecting a very low potential barrier of 0.1 along the projection normal direction and a very high potential barrier of 10.0 perpendicularly to restrict the quantum tunneling direction. Through adiabatic evolution and global energy minimization optimization calculations, the system automatically outputs the final collision-free arrangement coordinates of the primitives. Before drawing generation, the system also injects a 50mm hoisting gap in virtual space for reverse assembly, verifies the reserved rebar and sleeve with a 5mm alignment tolerance, and compares the two-dimensional and three-dimensional dimensions with a 2mm error threshold. If a collision or exceeding the tolerance is detected, an immediate interception warning is issued. Table 1 below shows the comparison data between the application of this invention and the traditional manual method with conventional BIM plugins during the four-month construction of the core fermentation chamber: Table 1. Comparison of the overall performance of the present invention and traditional methods; Based on the comparative data shown in Table 1, it can be seen that the BIM-based automatic disassembly and drawing system for prefabricated components proposed in this invention has significant performance advantages over traditional methods in the intelligent construction of brewing workshops, especially in terms of template utilization rate, element collision avoidance rate, drawing efficiency, dimensional accuracy and rework control, all of which have achieved comprehensive improvement.

[0034] Regarding template utilization, this invention maintains a high level of over 79% in all four typical component scenarios, far exceeding the average utilization rate of approximately 54% for traditional methods. For example, in the "ground cylinder protective cover module" scenario, traditional methods rely on manual experience for boundary segmentation, making it difficult to balance mold size and multiple physical constraints, easily generating waste materials at the edges, achieving only a utilization rate of 48.3%. In contrast, this invention achieves a utilization rate of 79.8% through multi-constraint parameterized matching and dynamic boundary optimization mechanisms, significantly reducing the cost of steel template mold making.

[0035] In terms of intelligent drawing generation and element collision avoidance, this invention introduces an improved quantum annealing algorithm to construct an anisotropic potential field, achieving a globally optimal solution for the arrangement of elements in the drawing, with a collision avoidance rate exceeding 97%. Traditional methods, due to the dense overlap of dimension lines and embedded part symbols, achieve a collision avoidance rate of only about 53%, requiring significant manual repositioning. This invention reduces the average drawing generation time per sheet from approximately 55 minutes to 8 minutes, a nearly seven-fold speedup, greatly enhancing design productivity.

[0036] It also demonstrates significant advantages in dimensional accuracy and rework control. Through virtual reverse assembly and a closed-loop verification mechanism, this invention controls the dimensional deviation rate in two and three dimensions to within 0.4%, reducing the on-site assembly rework rate to approximately 0.6%. Compared to the traditional method's nearly 5.5% dimensional deviation and 6.6% rework rate, this significantly reduces on-site cutting and rework issues such as sleeve misalignment and inability to insert reinforcing bars. The design-to-delivery cycle is also shortened from an average of over 14 days to approximately 3 days, ensuring high-precision, seamless installation of equipment such as the RGV intelligent vehicle.

[0037] Overall, this invention achieves efficient, accurate, and zero-collision drawing and delivery in key component scenarios such as track support platforms, partition walls, protective covers, and equipment foundations through multi-physical constraint decomposition, quantum annealing collision avoidance arrangement, and reverse assembly verification. It fills the technical shortcomings of traditional methods in the construction of large-scale intelligent brewing workshops, such as unreasonable decomposition and frequent drawing collisions.

[0038] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A BIM-based automatic disassembly and drawing generation system for prefabricated building components, characterized in that, Includes the following modules: The parameter extraction module is used to extract physical constraint parameters from the BIM model, match and bind geometric information with physical constraints, and obtain a standardized parameterized basic dataset. The boundary optimization module is used to select the splicing points and structural seams that are permissible under stress as the reference division points based on the normalized parameterized basic dataset, to perform preliminary region division, adjust the region boundaries, and obtain the optimal split boundary layout map. The sub-component generation module is used to identify connection relationships and disconnect cross-boundary steel bars and pipelines according to the optimal split boundary layout diagram, supplement and generate reserved anchorage ends or sleeve connection nodes, and at the same time assign the attribute information of the original component to the newly split sub-component and give it an independent identity identifier, thus obtaining the independently split sub-component. The intelligent drawing module is used to generate two-dimensional views based on independently split sub-components. It introduces an improved quantum annealing algorithm to construct the Hamiltonian of the anisotropic potential field modulation problem based on the projection normal. Through optimization calculation of global energy value minimization, it automatically detects and avoids overlapping collisions to obtain two-dimensional engineering drawings. The consistency verification module is used to reverse assemble all independent sub-components in virtual space according to their original positions and reserved hoisting gaps before outputting two-dimensional engineering drawings. If a collision or size discrepancy is found, the drawing will be intercepted and the error position will be indicated. The verification results will be obtained from the drawings and sub-components that have passed the verification. The manufacturing delivery module is used to package the verified drawings and the 3D models, 2D drawings and processing attributes of the sub-components, export the production data package, generate a tracking QR code, and obtain the processing dataset.

2. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 1, characterized in that, The parameter extraction module specifically includes: By calling the data interface of the BIM model, the three-dimensional solid objects in the model are traversed, and the length, width and height values ​​of each solid object are read as geometric dimensions. At the same time, the material attribute text configured for each solid object is read as material information. Extract the preset structural stress limits, preset mold table dimensions, preset transportation width and height limits, and preset component weight thresholds, and aggregate them into physical constraint parameters; The length, width, and height values ​​in the read geometric dimensions are compared with the preset mold table size and preset transportation width and height limits in the physical constraint parameters. When the geometric dimension is less than or equal to the preset mold table size and less than or equal to the preset transportation width and height limits, it is determined that the dimensions match. Simultaneously, the estimated weight of the component is obtained by multiplying the material density parameter corresponding to the material information with the volume calculated from the geometric dimensions. The estimated weight of the component is then compared with the preset component weight threshold in the physical constraint parameters. When the estimated weight of the component is less than or equal to the preset component weight threshold, it is determined that the weight is matched. The allowable design stress corresponding to the material information is then compared with the preset structural stress limit in the physical constraint parameters. When the allowable design stress is less than or equal to the preset structural stress limit, it is determined that the stress is matched. The geometric dimensions and material information that are determined to be size matching, weight matching, and stress matching are merged with the corresponding physical constraint parameters such as structural stress limits, mold size, transportation width and height limits, and component weight thresholds and written into the same dataset. The geometric information and physical constraint conditions are matched and bound to obtain a standardized parameterized basic dataset.

3. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 1, characterized in that, The boundary optimization module specifically includes: Based on the standardized parametric basic dataset, non-primary stress areas with design allowable stress less than or equal to the structural stress limit are found in the BIM model. Structural connection joints within the non-primary stress areas are extracted as structural joints, and splicing points and structural joints within the non-primary stress areas are marked as reference division points. Using the marked reference division points as cutting positions, the building space corresponding to the BIM model is divided along the horizontal and vertical directions, dividing the building space into multiple continuous preliminary area blocks, thus completing the preliminary area division of the building space. The weight of the region block is obtained by multiplying the volume calculated from the geometric dimensions of each preliminary region block by the material density parameter corresponding to the material information. The outer contour length and width parameters of each preliminary region block are extracted. The number of times the same outer contour length and width parameters appear is counted and divided by the total number of preliminary region blocks to obtain the template reuse rate. Calculate the weight of the current preliminary region block and the weight of the adjacent preliminary region block respectively. Subtract the weight of the adjacent preliminary region block from the weight of the current preliminary region block to obtain the weight difference. Determine whether the weight of the current preliminary region block is greater than the component weight threshold in the normalized parameterized basic dataset. If it is, move the corresponding reference division point towards the adjacent preliminary region block with a positive weight difference to reduce the geometric size of the current preliminary region block. Otherwise, keep the position of the current reference division point unchanged. Determine whether the template reuse rate of the current preliminary region block is lower than the preset utilization rate threshold. If so, modify the outer contour length and width parameters of the current preliminary region block to match the outer contour length and width parameters that appear most frequently in the statistics. Otherwise, keep the outer contour length and width parameters of the current preliminary region block unchanged. Repeatedly execute the steps of moving the reference division point and modifying the length and width parameters of the outer contour until the weight of all preliminary area blocks does not exceed the component weight threshold and the template reuse rate of all preliminary area blocks is not lower than the preset utilization rate threshold. Then, combine the adjusted area boundaries and reference division points to output the optimal split boundary layout diagram.

4. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 1, characterized in that, The sub-component generation module specifically includes: Based on the baseline division points in the optimal split boundary layout diagram and the adjusted area boundary, a split section is generated in the BIM model along the adjusted area boundary. The three-dimensional solid objects intersecting with the split section are extracted as each solid component. The coordinates of the intersection points of the split section with the geometric center lines of the steel bars and pipelines inside the solid components are calculated. The steel bar attributes and pipeline attributes on both sides of the intersection point coordinates are read and recorded to obtain the connection relationship between the steel bars and pipelines across the components. The steel bars whose geometric center lines intersect with the cutting section are marked as cross-boundary steel bars, and the pipelines whose geometric center lines intersect with the cutting section are marked as cross-boundary pipelines. At the coordinates of the intersection point between the cross-boundary steel bars and the cross-boundary pipelines and the cutting section, the three-dimensional geometric models of the cross-boundary steel bars and the cross-boundary pipelines are broken, so that the cross-boundary steel bars and the cross-boundary pipelines are separated into two independent parts at the intersection point coordinates. At the endpoint where the cross-boundary reinforcement is broken, the diameter and material information of the cross-boundary reinforcement are read. When the diameter of the cross-boundary reinforcement is greater than or equal to the preset diameter threshold, a sleeve connection node with the same diameter as the cross-boundary reinforcement is generated on the outside of the endpoint. When the diameter of the cross-boundary reinforcement is less than the preset diameter threshold, an outwardly extending reinforcement segment is generated on the outside of the endpoint as a reserved anchoring end. At the endpoint where the cross-boundary pipeline is broken, the cross-sectional dimension of the cross-boundary pipeline is read, and a pipeline interface with the same cross-sectional dimension as the cross-boundary pipeline is generated on the outside of the endpoint as a pipeline connection node. Extract the geometric dimensions and material information recorded in the original solid component before disconnection. Copy the geometric dimensions and material information and write it into the attribute data table of each newly disconnected sub-component. Generate a unique alphanumeric string for each newly split sub-component according to the preset encoding rules. Write the string into the attribute data table of the sub-component as an independent identifier, thus obtaining the independently split sub-component.

5. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 1, characterized in that, The intelligent image output module specifically includes: The normal vectors of the outer contour surfaces are extracted from the three-dimensional geometric models of the independently decomposed sub-components. The area value of each outer contour surface is multiplied by the corresponding normal vector, and the sum is divided by the total area of ​​all outer contour surfaces to calculate the area-weighted average normal vector. The direction perpendicular to the area-weighted average normal vector is determined as the main force-bearing surface direction of the independent sub-component. The angle perpendicular to the main force-bearing surface direction and with the largest projected area is calculated as the optimal viewing angle. The independent sub-component is orthogonally projected along the optimal viewing angle to generate a two-dimensional view. On the 3D geometric model of the independently split sub-components, extract the coordinates of all vertices, the midpoint coordinates of each edge line, and the endpoint coordinates of the arc curves. Summarize these coordinates and calculate them as the coordinates of the key points of the 3D model. Extract the coordinates of the leader points of the dimension annotations on the drawings of the independently split sub-components in the 2D view. Establish a one-to-one mapping link relationship between the leader point coordinates and the corresponding 3D model key point coordinates and bind them. When the coordinates of the 3D model key points are modified, calculate the coordinate difference synchronously according to the link relationship and update the value of the dimension annotation. Extract the bounding rectangles of the dimension lines, rebar markings, and embedded part symbols on the drawing. Map each bounding rectangle to a discrete state variable. Define the x-coordinate and y-coordinate of the center point of each discrete state variable and the length and width parameters of the bounding rectangle as the position state of the qubit. Use the initial arrangement of the dimension lines, rebar markings, and embedded part symbols in the two-dimensional view as the initial position state of the qubit. An improved quantum annealing algorithm is introduced to calculate the projection direction vector of the optimal viewing angle of the two-dimensional view as the normal vector generated by the projection. The direction of the parallel component of each discrete state variable along the normal vector generated by the projection is defined as the tunneling allowable direction, and the direction of the component perpendicular to the normal vector generated by the projection is defined as the tunneling restrictive direction. The initial Hamiltonian is calculated, and the problem Hamiltonian is calculated using the local potential field strength and the interaction potential field strength. The anisotropic potential field modulation problem Hamiltonian based on the projection normal is constructed to obtain the total Hamiltonian of the system. The optimization calculation is performed in the continuous evolution step to output the final position coordinates of all discrete state variables. The dimension lines, rebar markings, and embedded part symbols are rearranged in the 2D view according to the final position coordinates. The system automatically detects and avoids overlapping and collisions between dimension lines, rebar markings, and embedded part symbols to produce 2D engineering drawings.

6. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 5, characterized in that, The process involves calculating the initial Hamiltonian, using the local potential field strength and the interaction potential field strength to calculate the problem Hamiltonian, constructing the anisotropic potential field modulation problem Hamiltonian based on the projection normal, obtaining the total system Hamiltonian, and then performing optimization calculations in the continuous evolution step to output the final position coordinates of all discrete state variables. Specifically, this includes: The initial Hamiltonian is obtained by multiplying the transverse tunneling amplitude along the permitted tunneling direction with the preset transverse Pauli matrix operator of the corresponding discrete state variable, adding the transverse tunneling amplitude along the restricted tunneling direction with the preset longitudinal Pauli matrix operator of the corresponding discrete state variable, and summing the product results of all discrete state variables. The local potential field strength is obtained by multiplying the reciprocal of the shortest distance from the corresponding discrete state variable to the outer contour of the independent split sub-component with the preset attraction coefficient. The overlapping area is obtained by multiplying the overlap length of the bounding rectangles of every two discrete state variables on the horizontal and vertical axes. When the overlapping area is greater than zero, the interaction potential field strength is set to a preset positive penalty value. When the overlapping area is equal to zero, the interaction potential field strength is set to zero. The Hamiltonian of the problem is obtained by multiplying the local potential field strength of each discrete state variable with the corresponding preset Pauli Z operator and summing them, and by multiplying the interaction potential field strength between each pair of discrete state variables with the corresponding two preset Pauli Z operators and summing them. This constructs the Hamiltonian of the anisotropic potential field modulation problem based on the projection normal. Based on the normal vector generated by the projection, a directional potential field is injected. The lateral tunneling amplitude in the tunneling allowable direction is set to an extremely low barrier value, and the lateral tunneling amplitude in the tunneling restricted direction is set to an extremely high barrier value. This makes the barrier of the primitive along the projection normal direction extremely low and the barrier in the vertical direction extremely high, thus restricting the tunneling direction. The initial Hamiltonian is added to the problem Hamiltonian to obtain the total Hamiltonian of the system. An adiabatic evolution operation is performed on the total Hamiltonian of the system, and the transverse tunneling amplitude is gradually reduced to zero by multiplying the initial maximum value by a preset attenuation coefficient with the time step. At the same time, the local potential field strength and the interaction potential field strength are gradually increased to the target maximum value by multiplying the preset growth coefficient from zero. In each evolution step, the position and state of all current discrete state variables are combined into a column vector. The matrix corresponding to the total Hamiltonian of the system is multiplied by the column vector. The sum of squares of each element in the resulting column vector is calculated as the feature energy value of the current step. The feature energy value of the current step is subtracted from the feature energy value of the previous step, and the absolute value is taken as the energy difference. The energy difference is compared with a preset convergence threshold. When the energy difference is less than the preset convergence threshold, the evolution stops. The position state of all discrete state variables is read. The current feature energy value is compared with the lowest feature energy value in the history. If the current feature energy value is less than the lowest feature energy value in the history, the current position state is overwritten and saved as the final position coordinates. The final position coordinates of all discrete state variables are calculated and output through global energy minimization optimization.

7. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 1, characterized in that, The consistency verification module specifically includes: Read the three-dimensional geometric model and corresponding independent identity of all independent sub-components in the virtual space, extract the original coordinate position of each independent sub-component recorded before disassembly as the original position, and inject a gap of preset width between the outer contours of all independent sub-components as a reserved hoisting gap. Each independent sub-component is moved and assembled in reverse in virtual space according to its original position and the reserved hoisting gap. The three-dimensional geometric model surfaces of adjacent independent sub-components after reverse movement and assembly are extracted, the intersection volume between adjacent surfaces is calculated, and a geometric collision is determined when the intersection volume is greater than zero. The position where the intersection volume is greater than zero is recorded as the geometric collision position. In the independent disassembled sub-components after reverse movement assembly, read the three-dimensional coordinates of the reserved anchorage end and the sleeve connection node generated at the break of the cross-boundary reinforcement, calculate the spatial distance between the center point of the reserved anchorage end and the center point of the corresponding sleeve connection node, and determine that the reserved reinforcement and the sleeve are not aligned when the spatial distance is greater than the preset alignment tolerance, and record the position where the spatial distance is greater than the preset alignment tolerance. Read the three-dimensional coordinates of the pipeline connection node and the corresponding pre-embedded junction box interface generated at the cross-border pipeline break point, calculate the spatial distance between the center point of the pipeline connection node and the center point of the pre-embedded junction box interface, and determine that the reserved pipeline and the pre-embedded junction box are not aligned when the spatial distance is greater than the preset alignment tolerance, and record the position where the spatial distance is greater than the preset alignment tolerance. Extract the values ​​of all dimensions in the 2D engineering drawings as 2D dimension data, read the coordinates of the corresponding key points in the 3D model and calculate the spatial distance between the two points as 3D dimension data, subtract the 2D dimension data and 3D dimension data at the same location and take the absolute value. When the absolute value is greater than the preset dimension error threshold, it is determined that the dimensions of the 2D engineering drawings and the dimensions of the 3D model are inconsistent, and the position of the dimension annotation corresponding to the absolute value being greater than the preset dimension error threshold is recorded. The system summarizes the locations of geometric collisions, misalignments between reserved reinforcing bars and sleeves, misalignments between reserved pipelines and embedded junction boxes, and dimensions inconsistencies between the dimensions on the 2D engineering drawings and the dimensions on the 3D model as error locations. It intercepts and outputs the 2D engineering drawings and prompts the error locations. If no error is found, it is determined that the verification is correct and the verification of the drawings and sub-components is passed.

8. The BIM-based automatic disassembly and drawing system for prefabricated building components according to claim 1, characterized in that, The manufacturing delivery module specifically includes: The verified drawings, the 3D model of the sub-component, the 2D drawings, and the processing attributes are merged into the same file and packaged. According to the preset data format requirements of the factory production system, the processing attributes are converted into a material cutting table, the steel bar outline in the 2D drawings is converted into steel bar processing data, and the processing trajectory of the 3D model is converted into CNC machine tool processing code. The material cutting list, rebar processing data, and CNC machine tool processing code are merged and exported as a production data package with one click. The unique identifier of each independent sub-component is read and converted into a QR code image. The QR code image is then attached to the corresponding production data package to obtain the processing dataset.

Images