Collision avoidance and multi-objective optimization method and device for steel bar spatial arrangement

Abstract

The invention discloses a collision avoidance and multi-objective optimization method and device for steel bar spatial arrangement, and relates to the technical field of digital design and intelligent construction of building structures. The method comprises the following steps: firstly, constructing a parameterized three-dimensional reinforcement arrangement environment comprising a concrete host domain, an obstacle domain and a protective layer; a multi-level collision detection strategy based on spatial indexes is adopted, and interference objects are quickly locked through bounding box coarse screening and geometric accurate intersection; establishing a multi-target evaluation system fusing standard hard constraints and construction reachability soft constraints; an improved multi-objective optimization algorithm introducing a physical collision feedback mechanism is utilized, and a collision detection result is converted into physical repulsive force to directly drive reinforcing steel bar position updating; and finally, numerical control machining data meeting the BVBS standard is generated. According to the method, the algorithm convergence efficiency is remarkably improved through a physical feedback mechanism, and collision-free automatic arrangement of the reinforcing steel bars meeting the construction and machining requirements under the complex special-shaped structure is achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction technology in civil engineering, specifically to a collision avoidance and multi-objective optimization method and apparatus for the spatial arrangement of reinforced concrete. Background Technology

[0002] With the acceleration of industrialization in construction, rebar layout is gradually shifting from two-dimensional plan drawings to three-dimensional digital models. Collision checking using BIM software has become a common practice. For example, patent CN114329693B identifies conflicts during rebar detailing by performing Boolean operations on the solid model; patent CN113032861A proposes a component collision analysis method based on a BIM model. These technologies can effectively identify interference between rebars and between rebars and embedded parts, but they are usually limited to outputting collision locations; subsequent avoidance adjustments still rely on manual work by designers. In cases of complex nodes or varied component shapes, manual correction is labor-intensive, and local modifications may lead to new interferences. On the other hand, there are also automated rebar layout approaches based on rule bases in the industry. For example, patent CN104156544A automatically generates beam-column node rebar through a pre-set three-dimensional node module. Such methods are efficient for regular components, but due to the limitations of the rule base itself, they often fail to cover actual needs when encountering irregular curved surfaces, non-standard intersecting nodes, or scenarios with multiple obstacles, requiring manual intervention for adjustments. Furthermore, some studies have attempted to improve the automation level of reinforcement placement using metaheuristic algorithms. For example, patent number CN114611191A uses an improved genetic algorithm for structural optimization design, and related academic work has also explored the use of particle swarm optimization to handle reinforcement congestion problems. These algorithms can handle multi-objective optimization to some extent, but their iterative process lacks direct awareness of the geometric environment, typically relying on random search, leading to slow convergence, a tendency to get trapped in local optima, and sometimes resulting in reinforcement shapes that fail to meet workability or construction accessibility requirements.

[0003] Under the aforementioned technical conditions, the detailed design of reinforcing bars in complex components still faces significant bottlenecks. For high-density reinforcement areas, traditional optimization algorithms often struggle to find a collision-free layout that meets code requirements within a limited timeframe. Iterative processes are prone to computational stagnation or repeated adjustments that fail to completely eliminate interference. As component geometry becomes increasingly complex, solution methods relying solely on penalty functions or random searches are insufficient to effectively drive solutions towards feasible regions. Even if a geometrically "collision-free" layout is achieved, existing methods lack effective guarantees regarding constructability. Many optimization results focus only on spatial avoidance relationships, failing to simultaneously consider construction factors such as rebar processing radius, frequency of non-standard bends, and node operation space. In actual construction, excessive bend angles or insufficient operating surfaces can render the design unfeasible, leading to rework during the construction phase. Simultaneously, data disconnect persists between the design model and processing equipment in the current workflow. Reinforcing bar layout results often require manual re-processing into a format suitable for CNC machining; this manual conversion is time-consuming and prone to errors in complex components. For the pursuit of an integrated process of digital construction, how to make the layout results directly correspond to standardized processing instructions is also a challenge that existing technologies have not yet solved.

[0004] In summary, existing technologies still suffer from insufficient automation, low collision avoidance efficiency, and difficulty in balancing structural specifications and construction techniques when dealing with complex spatial conditions, high-density reinforcement, and components with multiple types of obstacles. How to effectively obtain geometric feedback such as collision location and depth in a 3D environment and directly apply it to the optimization process to improve solution efficiency and layout quality remains an unsolved technical challenge in the industry. Summary of the Invention

[0005] The purpose of this invention is to provide a collision avoidance and multi-objective optimization method and apparatus for the spatial arrangement of reinforced concrete, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a collision avoidance and multi-objective optimization method for the spatial arrangement of reinforcing bars, comprising the following steps: S1: Construct a parametric reinforcement environment: Obtain the geometric information of the concrete component, determine the host domain of the reinforcement layout, identify the obstacle domain formed by embedded parts and holes, and offset the host domain boundary inward according to the protective layer thickness required by the design specifications to form a three-dimensional effective space in which the reinforcement centerline is allowed to exist. S2: Generate initial steel reinforcement population: Generate an initial steel reinforcement population within the effective space. Each individual steel reinforcement is stored in a parameterized form, including spatial position coordinates, axial direction vector and cross-sectional geometric parameters. S3: Perform multi-level collision detection: Perform collision detection on the rebar population. After coarse screening using bounding box spatial indexing, perform geometrically precise intersection operations on potential interference objects, and output the collision state information of each rebar. The collision state information includes the collision depth scalar d and the normal unit vector at the collision point. ; S4: Establish a multi-objective evaluation system: Construct an evaluation function that includes mandatory engineering constraints and non-mandatory optimization objectives, where mandatory engineering constraints serve as the feasibility criterion for the solution, and non-mandatory optimization objectives serve as the basis for fitness calculation; S5: Perform iterative optimization with physical collision feedback: Run a multi-objective optimization algorithm for iterative solution. In each iteration, based on the collision depth d and normal unit vector obtained in S3... Calculate the physical repulsive force vector The physical repulsive force vector The vector superposition with the original position update components of the optimization algorithm forms the actual position update amount of the individual steel bars, enabling the steel bars to evolve towards the collision-free region under the premise of satisfying the non-mandatory optimization objective, until the preset convergence condition is met; S6: Output reinforcement layout results: The optimized and converged reinforcement geometry is rationalized, and the reinforcement layout result data is output.

[0007] Furthermore, the physical repulsive force vector in S5 The calculation formula is: ; In the formula: This is the repulsive force weighting coefficient, which decreases as the number of iterations increases; is the hardness coefficient of the colliding object; d is the scalar value of the collision depth. The normal unit vector pointing towards the repelled rebar at the point of collision; and the stiffness coefficient of the fixed obstacle. Greater than the hardness coefficient between individual steel bars; Among them, the repulsive force weighting coefficient Use any of the following methods with the number of iterations Decreasing: (1) Linear decay: ; In the formula, These are the initial weighting coefficients. This represents the maximum number of iterations. (2) Exponential decay: ; In the formula, This is the decay rate coefficient, and its value range is... ; Stiffness coefficient of the colliding object The rules for determining the value are as follows: (1) When the collision object is a concrete boundary or obstacle domain , The range of values ​​is ; (2) When the collision object is another individual steel bar. , The range of values ​​is And satisfy .

[0008] Furthermore, the multi-level collision detection in S3 includes: Level 1: Spatial Index Coarse Screening: Construct an axial bounding box for each rebar, divide the space using an octree or hierarchical bounding box (BVH) structure, quickly eliminate rebar pairs that are unlikely to collide in space by judging the intersection of bounding boxes, and output a set of potential interference objects. Level 2: Geometric Precise Intersection: Performs precise geometric intersection operations on the steel reinforcement pairs in the set of potential interference objects. Based on the geometry of the steel reinforcement, it selects a line segment-line segment, line segment-plane, or cylinder-cylinder intersection algorithm to calculate the collision depth. and the normal unit vector at the point of collision .

[0009] Furthermore, when the concrete component is an irregular curved surface component, the method for generating the initial steel reinforcement population in S2 is as follows: extract the principal curvature direction of the host domain surface or generate a streamline field according to the principal stress distribution of the structure, initialize the spatial topology and orientation of the individual steel reinforcements along the streamline field direction, so that the initial steel reinforcement arrangement conforms to the curved surface shape of the component.

[0010] Furthermore, mandatory engineering constraints in S4 include: Protective layer thickness constraint: The distance from the centerline of the reinforcing bar to the concrete boundary shall not be less than the designed protective layer thickness; Reinforcing bar clear spacing constraint: The minimum clear spacing between adjacent reinforcing bar surfaces shall not be less than the value specified in the code; Anchorage length constraint: The anchorage length at the end of the reinforcing bar shall meet the structural design requirements; Obstacle safety distance constraint: The minimum distance between the reinforcing bar and the boundary of the obstacle area shall not be less than the preset safety margin; Non-mandatory optimization objectives include at least one of the following: Construction accessibility: minimizing the interference volume between the virtual operation sphere and the reinforcing steel at key nodes as the evaluation index; Processing complexity: minimizing the occurrence of non-standard bending angles and the total number of bending times of the reinforcing steel as the evaluation index; Material consumption: minimizing the total length of the reinforcing steel as the evaluation index.

[0011] Furthermore, the specific methods in S6 include: (1) Geometric rationalization: The optimized centerline curve of the steel bar is fitted as a combination of straight line segments and circular arc segments, and the bending angle is corrected. When the deviation of the bending angle from 90° or 135° is less than the preset angle tolerance, it is forcibly corrected to the standard angle, and the minimum bending radius is constrained according to the steel bar diameter. (2) Data output: The geometric parameters, material information and bending sequence of the steel bars are parsed into instruction formats that conform to the CNC machining equipment standards. The instruction formats include BVBS format or IFC format.

[0012] Furthermore, in S5, during the iterative optimization process, when it is detected that the collision depth of a certain rebar has not decreased significantly for N consecutive generations, a topology mutation operation is performed on the rebar. The topology mutation operation includes inserting control points on the center line of the rebar or adjusting the anchorage method at the end of the rebar to increase the geometric degrees of freedom of the rebar.

[0013] Furthermore, the construction of the obstacle domain in S1 includes: obtaining the dynamic positioning path of the component during the assembly process, scanning the reserved reinforcing bars of the lower structure in reverse along the positioning path, generating the reinforcing bar scanning body, and incorporating the reinforcing bar scanning body into the obstacle domain to ensure that the optimized reinforcement arrangement does not interfere with the assembly process.

[0014] This invention provides another technical solution: a collision avoidance and multi-objective optimization device for the spatial arrangement of reinforcing bars, comprising: Memory, used to store computer programs; The processor, coupled to the memory, is used to execute a computer program to implement the functions of the following modules: an environment construction module for constructing a parameterized reinforcement environment including a host domain, an obstacle domain, and a protective layer boundary; a population initialization module for generating an initial reinforcement population; and a collision detection module for performing multi-level collision detection and outputting the collision depth d and the normal vector. The evaluation system module is used to establish multi-objective evaluation functions; the optimization solution module is used to run multi-objective optimization algorithms based on the collision depth d and the normal vector. Calculate the physical repulsive force vector and the repulsive force vector The data is superimposed into the position update equation of the individual steel bars; the data output module is used to rationalize the optimized steel bar geometry and output the result data.

[0015] The present invention also provides a technical solution: a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of a collision avoidance and multi-objective optimization method for steel reinforcement spatial arrangement.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. The present invention provides a collision avoidance and multi-objective optimization method and apparatus for the spatial arrangement of reinforcing bars. It introduces a repulsive force feedback mechanism based on physical collision depth, which provides clear gradient direction guidance for the optimization algorithm, enabling the reinforcing bars to actively adjust along the shortest path to avoid collision. This effectively solves the problems of slow convergence and easy getting trapped in local optima in the traditional penalty function method, and significantly improves the convergence efficiency of the algorithm under high-dimensional constraints.

[0017] 2. The present invention provides a collision avoidance and multi-objective optimization method and apparatus for the spatial arrangement of reinforcing bars. By introducing NURBS surface reconstruction and streamline-guided reinforcement technology, the algorithm can generate a reinforcing bar topology that conforms to the main curvature direction of the component surface. It is suitable for irregular structures, ensures the rationality of the arrangement, and realizes the shape-adaptive arrangement of reinforcing bars in irregular and complex structures.

[0018] 3. The present invention provides a collision avoidance and multi-objective optimization method and device for the spatial arrangement of reinforcing bars. The optimized reinforcing bars undergo geometric rationalization and workability verification, and directly generate CNC machining instructions that conform to the BVBS standard, eliminating the data barrier between the design model and the reinforcing bar processing equipment.

[0019] 4. The present invention provides a collision avoidance and multi-objective optimization method and device for the spatial arrangement of reinforcing bars. The multi-objective evaluation system integrates soft constraints on construction accessibility and reserves operational space for key processes such as sleeve grouting and reinforcing bar connection in advance during the design stage, which significantly reduces the on-site change rate and rework cost, and takes into account both theoretical collision-free and on-site construction feasibility. Attached Figure Description

[0020] Figure 1 This is a diagram illustrating the overall technical architecture of the present invention; Figure 2 This is a schematic diagram of the parametric rebar modeling of the present invention; Figure 3 This is a schematic diagram of the multi-level collision detection of the present invention; Figure 4 This is a schematic diagram of the multi-objective evaluation method in this invention; Figure 5 This is the iterative graph of the multi-objective optimization algorithm in this invention; Figure 6 This is a schematic diagram of a complex beam-column node collision detection model in an embodiment of this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 The collision avoidance and multi-objective optimization method for steel reinforcement spatial arrangement provided in this embodiment of the invention includes the following steps: S1: Constructing a parametric reinforcement environment: Acquire the geometric information of the concrete component, determine the host domain for the reinforcement layout, identify embedded parts and holes constituting obstacle domains (such as embedded parts and holes), and offset the host domain boundary inward according to the protective layer thickness required by the design specifications to form a three-dimensional effective space where the reinforcement centerline is allowed to exist; among them, the construction of the obstacle domain includes: acquiring the dynamic positioning path of the component during the assembly process, scanning the reserved reinforcement bars of the lower structure in reverse along the positioning path, generating the reinforcement bar scan volume, and incorporating the reinforcement bar scan volume into the obstacle domain to ensure that the optimized reinforcement layout does not interfere with the assembly process. For irregular curved surface components, NURBS technology is used to reconstruct the component surface, such as... Figure 2 As shown.

[0023] S2: Generate initial reinforcement population: Generate initial reinforcement population within the effective space. Each individual reinforcement bar is stored in a parametric form, including spatial position coordinates, axial direction vector and cross-sectional geometric parameters. When the concrete component is an irregular curved surface component, the specific method for generating the initial reinforcement population is as follows: extract the principal curvature direction of the host domain surface or generate a streamline field according to the principal stress distribution of the structure. Initialize the spatial topology and orientation of the individual reinforcement bars along the streamline field direction so that the initial reinforcement arrangement conforms to the curved surface shape of the component.

[0024] S3: Perform multi-level collision detection: Perform collision detection on the rebar population. After coarse screening using bounding box spatial indexing, perform geometrically precise intersection operations on potential interference objects, and output the collision state information of each rebar. The collision state information includes the collision depth scalar d and the normal unit vector at the collision point. Among them, multi-level collision detection adopts a two-level detection strategy of "coarse screening-fine calculation", such as... Figure 3 As shown: Level 1 (Coarse Screening): Spatial Index Coarse Screening: Constructs an axial bounding box for each rebar, first filtering out rebar pairs that will not collide. Spatial partitioning is performed using an octree or hierarchical bounding box (BVH) structure. Rebar pairs that are unlikely to collide in space are quickly eliminated by judging the intersection of bounding boxes, outputting a set of potential interference objects. Level 2 (Precise Calculation): Geometric Precision Intersection: Perform precise geometric intersection operations on the steel reinforcement pairs in the set of potential interference objects. Based on the geometry of the steel reinforcement, select a line segment-line segment, line segment-plane, or cylinder-cylinder intersection algorithm to calculate the collision depth. and the normal unit vector at the point of collision .

[0025] S4: Establish a multi-objective evaluation system: Construct an evaluation function that includes mandatory engineering constraints and non-mandatory optimization objectives. Mandatory engineering constraints (such as concrete cover thickness, rebar clearance, anchorage length, and obstacle safety distance) serve as feasibility criteria for solutions, while non-mandatory optimization objectives (such as construction accessibility, processing complexity, and material usage) serve as the basis for fitness calculations. Specifically, the concrete cover thickness constraint requires that the distance from the rebar centerline to the concrete boundary be no less than the designed concrete cover thickness; the rebar clearance constraint requires that the minimum clearance between adjacent rebar surfaces be no less than the value specified in the code; the anchorage length constraint requires that the anchorage length at the rebar end meet the structural design requirements; and the obstacle safety distance constraint requires that the minimum distance between the rebar and the boundary of the obstacle domain be no less than the preset safety margin. In the non-mandatory optimization objectives, construction accessibility is evaluated by minimizing the interference volume between the virtual operation sphere and the reinforcing steel at key nodes; processing complexity is evaluated by minimizing the number of non-standard bending angles and the total number of reinforcing steel bends; and material consumption is evaluated by minimizing the total length of the reinforcing steel. Figure 4 As shown; S5: Perform iterative optimization with physical collision feedback: Run a multi-objective optimization algorithm for iterative solution. During the iterative optimization process, when it is detected that the collision depth of a certain rebar has not decreased significantly for N consecutive generations, perform a topology mutation operation on that rebar. The topology mutation operation includes inserting control points on the centerline of the rebar or adjusting the anchorage method at the end of the rebar to increase the geometric degrees of freedom of the rebar, such as... Figure 5 As shown: In each iteration, based on the collision depth d and normal unit vector obtained from S3... Calculate the physical repulsive force vector : ; In the formula: This is the repulsive force weighting coefficient, which decreases as the number of iterations increases; is the hardness coefficient of the colliding object; d is the scalar value of the collision depth. The normal unit vector pointing towards the repelled rebar at the point of collision; and the stiffness coefficient of the fixed obstacle. Greater than the hardness coefficient between individual steel bars; Among them, the repulsive force weighting coefficient Use any of the following methods with the number of iterations Decreasing: (1) Linear decay: ; In the formula, The weight coefficients decrease with the number of iterations. These are the initial weighting coefficients. This represents the maximum number of iterations. (2) Exponential decay: ; In the formula, This is the decay rate coefficient, and its value range is... ; Stiffness coefficient of the colliding object The rules for determining the value are as follows: (1) When the collision object is a concrete boundary or obstacle domain , The range of values ​​is ; (2) When the collision object is another individual steel bar. , The range of values ​​is And satisfy

[0026] The above physical repulsion force vector The vector superposition with the original position update components of the optimization algorithm forms the actual position update amount of the individual steel bars, enabling the steel bars to evolve towards the collision-free region under the premise of satisfying the non-mandatory optimization objective, until the preset convergence condition is met; S6: Output Reinforcement Layout Results: This function optimizes the reinforcement geometry after convergence and outputs the reinforcement layout results, including: (1) Geometric rationalization: The optimized centerline curve of the steel bar is fitted as a combination of straight line segments and circular arc segments, and the bending angle is corrected. When the deviation of the bending angle from 90° or 135° is less than the preset angle tolerance, it is forcibly corrected to the standard angle, and the minimum bending radius is constrained according to the steel bar diameter. (2) Data output: The geometric parameters, material information and bending sequence of the steel bars are parsed into instruction formats that conform to the CNC machining equipment standards. The instruction formats include BVBS format or IFC format.

[0027] To further explain and illustrate the present invention, the following specific implementation examples are also provided: Case 1: Obstacle avoidance optimization for high-density reinforced beam-column joints in cast-in-place frame structures, such as... Figure 6 As shown: This is used to verify the convergence performance and obstacle avoidance effect of the method of the present invention under strong constraints. 1. Verification Scenario Construction: A beam-column joint in the transfer layer of a frame structure is selected. The joint includes 8 circular concrete columns with a diameter of 1200mm and 12 frame beams with a cross-section of 800*1500mm. Four DN200 rainwater downpipes and two DN1500 condensate pipes pass through the core area of ​​the joint. The reinforcement density (volume ratio) of the joint is 4.2%.

[0028] 2. Parametric reinforcement layout environment and initial state: S1 (Environment Construction): Set the concrete protective layer thickness C=30mm. Create a cylindrical model for the pipe crossing the node, and establish a "barrier domain" extending 20mm outwards from its surface. .

[0029] S2 (Population Initialization): Generates an initial population of 864 steel bars within the host domain. The initial state contains 156 collision pairs, with a maximum collision depth of 25mm.

[0030] 3. Spatial Indexing and Collision Detection Configuration: S3 (Collision Detection): Constructing an octree spatial index. The detection process employs a "coarse screening-fine calculation" strategy. Through the indexing mechanism, the computational cost of performing precise intersection calculations is theoretically reduced from 3.7 * 10^6. 5 The number of times decreased to 2847.

[0031] 4. Optimization solution incorporating physical feedback: Steps S4-S5 (Iterative Optimization): An improved multi-objective particle swarm optimization (MOPSO) algorithm is employed. Maximum number of iterations. Simultaneously, a physical repulsion mechanism is used, and a repulsion weighting coefficient is set. The number of iterations t decreases linearly, and the initial value is... Obstacle hardness coefficient Hardness coefficient between steel bars .

[0032] Calculation Example (Generation 5): Collision between longitudinal reinforcement and rainwater pipe, collision depth , legal direction .

[0033] Weighting coefficient .

[0034] Repulsive force vector The calculation is as follows:

[0035] This force drives the reinforcing bar to generate approximately [a certain amount] along the normal direction. The displacement.

[0036] 5. Optimization Results and Comparative Verification: After multiple generations of iterative calculations, the algorithm reached convergence. In the final solution, the number of collision pairs between the reinforcing bars and the pipes is 0, and the minimum clear distance between the reinforcing bars meets the code requirements.

[0037] Compared with the traditional penalty function method, the method of this invention achieves complete collision-free operation after multiple iterations, reducing computation time by approximately [amount missing]. .

[0038] 6. Data Output: S6 (Rationalization and Output): Post-processes the optimized rebar geometry, forcibly correcting minor angular deviations. The final result is machining data conforming to BVBS format, which can be directly imported into CNC machining equipment.

[0039] Case 2: Optimization of constructability and generation of CNC data for prefabricated shear wall components, used to verify the optimization capability of this invention under dynamic assembly path constraints and soft constraints of construction operation space: 1. Verification Scenario Construction: Precast shear wall components are selected, with grouting sleeves and embedded diagonal supports integrated within the components. The technical challenge lies in meeting the dynamic avoidance requirements during hoisting and the operational space requirements for machinery during grouting operations.

[0040] 2. Definition of dynamic assembly path and obstacle domain: S1 (Environment Construction): Introduces a dynamic interference detection mechanism based on the scanning volume.

[0041] Obstacle domain synthesis: Scan the reserved reinforcing bars of the lower structure in reverse along the assembly path to generate the reinforcing bar scanning volume. Incorporate it into the barrier domain Ensure that the hoisting access is unobstructed.

[0042] 3. Quantification of soft constraints on construction operation space: S4 (Evaluation System): Define the radius at the grouting port and grouting outlet of each sleeve. The coefficient requirements for the "virtual operation sphere" are as follows: Hardness coefficient of hard obstacle (sleeve body) The repulsive force has a high weight.

[0043] Stiffness coefficient of soft obstacles (operating spheres) This generates a weaker repulsive force and adds a penalty term for "minimizing interference volume" to the objective function.

[0044] 4. Multi-objective optimization solution process: S5 (Iterative Optimization): Employing improved... algorithm.

[0045] After final convergence, the hard collision elimination is zero, and the interference volume of the manipulated sphere is reduced to 22 cm³. 3(A decrease of 96.8%), the number of non-standard bends was reduced to 0.

[0046] 5. Reinforcement geometry optimization and BVBS output: S6 (Result Output): Orthogonalizes the optimized reinforcing bars. Reinforcing bars with bending angles within the range of 90°±5° are forcibly corrected to a standard 90° bend. The final data is parsed into BVBS format, and the processing pass rate has been verified to be 100%.

[0047] Case 3: Streamline guidance and topology optimization for irregularly shaped reinforcing bars in the anchorage zone of a long-span cable-stayed bridge tower, used to verify the adaptability of this invention in complex curved surfaces and ultra-large-scale reinforcement layout scenarios: 1. Parametric Surface Reconstruction and Environment Definition: S1 (Environment Construction): Based on the surface node coordinates, construct a third-order NURBS surface on the outer surface of the anchorage zone. Construct a barrier domain containing the steel anchor box and prestressed ducts.

[0048] 2. Population initialization based on streamline field: S2 (Population Generation): Employs a streamlined guided initialization method. Constructs a hybrid vector field. ( In the direction of principal curvature (In the direction of principal tensile stress). Along The initial reinforcement centerline is generated, and the average radius of curvature is increased by 42% compared to the straight reinforcement layout.

[0049] 3. Large-scale collision detection and adaptive topology mutation: S3 (Collision Detection): For a scale of 4000+ steel bars, a hierarchical bounding box (BVH) is used for spatial indexing, which reduces the amount of calculation for precise intersection by 99.3%.

[0050] Step S5 (Iterative Optimization and Mutation): Introduce an adaptive topology mutation mechanism, wherein, Triggering condition: When the collision depth decrease rate of a certain rebar is less than 5% for 8 consecutive generations, it is determined to be "deadlock".

[0051] Mutation operation—Insert a new control point at the NURBS parameter position corresponding to the collision point and apply an offset along the collision normal.

[0052] 4. Optimization Result Verification: After 76 iterations, the algorithm fully converged. The streamlined field-based reinforcement scheme reduced the local stress concentration factor of the concrete in the anchorage zone from 1.78 to 1.42.

[0053] 5. Digital delivery: S6 (Data Output): Output: Export to IFC4.0 standard format.

[0054] Processing data: The Douglas-Peucker algorithm is used to discretize the NURBS curve into "straight lines". The "arc" combination generates BVBS data containing special shape codes, supporting 3D CNC bending machine processing.

[0055] In summary, this invention provides a collision avoidance and multi-objective optimization method and apparatus for the spatial arrangement of reinforcing bars. It constructs a parameterized three-dimensional reinforcement environment comprising a concrete host domain, an obstacle domain, and a protective layer; employs a multi-level collision detection strategy based on spatial indexing, rapidly identifying interference objects through bounding box coarse screening and geometrically precise intersection calculation; establishes a multi-objective evaluation system integrating hard constraints from specifications and soft constraints from construction accessibility; and utilizes an improved multi-objective optimization algorithm incorporating a physical collision feedback mechanism to transform collision detection results into physical repulsion forces that directly drive the updating of reinforcing bar positions; ultimately generating CNC machining data conforming to the BVBS standard. This invention significantly improves the algorithm's convergence efficiency through the physical feedback mechanism, achieving collision-free automatic reinforcement arrangement that meets construction and processing requirements under complex irregular structures.

[0056] Based on this, the present invention also provides another technical solution: a collision avoidance and multi-objective optimization device for the spatial arrangement of reinforcing bars, used to implement the above method, comprising: Memory, used to store computer programs; The processor, coupled to memory, executes computer programs to implement the functions of the following modules: an environment construction module for constructing a parameterized reinforcement environment including the host domain, obstacle domain, and protective layer boundaries; a population initialization module for generating an initial reinforcement population; and a collision detection module for performing multi-level collision detection and outputting the collision depth d and normal vector. The evaluation system module is used to establish multi-objective evaluation functions; the optimization solution module is used to run multi-objective optimization algorithms based on the collision depth d and the normal vector. Calculate the physical repulsive force vector and the repulsive force vector The data is superimposed into the position update equation of the individual steel bars; the data output module is used to rationalize the optimized steel bar geometry and output the result data.

[0057] 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 collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars, characterized in that, Includes the following steps: S1: Construct a parametric reinforcement environment: Obtain the geometric information of the concrete component, determine the host domain of the reinforcement layout, identify the obstacle domain formed by embedded parts and holes, and offset the host domain boundary inward according to the protective layer thickness required by the design specifications to form a three-dimensional effective space in which the reinforcement centerline is allowed to exist. S2: Generate initial steel reinforcement population: Generate an initial steel reinforcement population within the effective space. Each individual steel reinforcement is stored in a parameterized form, including spatial position coordinates, axial direction vector and cross-sectional geometric parameters. S3: Perform multi-level collision detection: Perform collision detection on the rebar population. After coarse screening using bounding box spatial indexing, perform geometrically precise intersection operations on potential interference objects, and output the collision state information of each rebar. The collision state information includes the collision depth scalar d and the normal unit vector at the collision point. ; S4: Establish a multi-objective evaluation system: Construct an evaluation function that includes mandatory engineering constraints and non-mandatory optimization objectives, where mandatory engineering constraints serve as the feasibility criterion for the solution, and non-mandatory optimization objectives serve as the basis for fitness calculation; S5: Perform iterative optimization with physical collision feedback: Run a multi-objective optimization algorithm for iterative solution. In each iteration, based on the collision depth d and normal unit vector obtained in S3... Calculate the physical repulsive force vector The physical repulsive force vector The vector superposition with the original position update components of the optimization algorithm forms the actual position update amount of the individual steel bars, enabling the steel bars to evolve towards the collision-free region under the premise of satisfying the non-mandatory optimization objective, until the preset convergence condition is met; S6: Output reinforcement layout results: The optimized and converged reinforcement geometry is rationalized, and the reinforcement layout result data is output.

2. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, Physical repulsive force vector in S5 The calculation formula is: ； In the formula: This is the repulsive force weighting coefficient, which decreases as the number of iterations increases; is the hardness coefficient of the colliding object; d is the scalar value of the collision depth. The normal unit vector pointing towards the repelled rebar at the point of collision; and the stiffness coefficient of the fixed obstacle. Greater than the hardness coefficient between individual steel bars; Among them, the repulsive force weighting coefficient Use any of the following methods with the number of iterations Decreasing: (1) Linear decay: ； In the formula, These are the initial weighting coefficients. This represents the maximum number of iterations. (2) Exponential decay: ； In the formula, This is the decay rate coefficient, and its value range is... ; Stiffness coefficient of the colliding object The rules for determining the value are as follows: (1) When the collision object is a concrete boundary or obstacle domain , The range of values ​​is ; (2) When the collision object is another individual steel bar. , The range of values ​​is And satisfy .

3. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, Multi-level collision detection in S3 includes: Level 1: Spatial Index Coarse Screening: Construct an axial bounding box for each rebar, divide the space using an octree or hierarchical bounding box (BVH) structure, quickly eliminate rebar pairs that are unlikely to collide in space by judging the intersection of bounding boxes, and output a set of potential interference objects. Level 2: Geometric Precise Intersection: Performs precise geometric intersection operations on the steel reinforcement pairs in the set of potential interference objects. Based on the geometry of the steel reinforcement, it selects a line segment-line segment, line segment-plane, or cylinder-cylinder intersection algorithm to calculate the collision depth. and the normal unit vector at the point of collision .

4. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, When the concrete component is an irregular curved surface component, the method for generating the initial steel reinforcement population in S2 is as follows: extract the principal curvature direction of the host domain surface or generate a streamline field according to the principal stress distribution of the structure, initialize the spatial topology and orientation of the individual steel reinforcements along the streamline field direction, so that the initial steel reinforcement arrangement conforms to the curved surface shape of the component.

5. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, Mandatory engineering constraints in S4 include: Protective layer thickness constraint: The distance from the centerline of the reinforcing bar to the concrete boundary shall not be less than the designed protective layer thickness; Reinforcing bar clear spacing constraint: The minimum clear spacing between adjacent reinforcing bar surfaces shall not be less than the value specified in the code; Anchorage length constraint: The anchorage length at the end of the reinforcing bar shall meet the structural design requirements; Obstacle safety distance constraint: The minimum distance between the reinforcing bar and the boundary of the obstacle area shall not be less than the preset safety margin; Non-mandatory optimization objectives include at least one of the following: Construction accessibility: minimizing the interference volume between the virtual operation sphere and the reinforcing steel at key nodes as the evaluation index; Processing complexity: minimizing the occurrence of non-standard bending angles and the total number of bending times of the reinforcing steel as the evaluation index; Material consumption: minimizing the total length of the reinforcing steel as the evaluation index.

6. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, The specific methods in S6 include: (1) Geometric rationalization: The optimized centerline curve of the steel bar is fitted as a combination of straight line segments and circular arc segments, and the bending angle is corrected. When the deviation of the bending angle from 90° or 135° is less than the preset angle tolerance, it is forcibly corrected to the standard angle, and the minimum bending radius is constrained according to the steel bar diameter. (2) Data output: The geometric parameters, material information and bending sequence of the steel bars are parsed into instruction formats that conform to the CNC machining equipment standards. The instruction formats include BVBS format or IFC format.

7. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, In S5, during the iterative optimization process, when it is detected that the collision depth of a certain rebar has not decreased significantly for N consecutive generations, a topology mutation operation is performed on the rebar. The topology mutation operation includes inserting control points on the center line of the rebar or adjusting the anchorage method at the end of the rebar to increase the geometric degrees of freedom of the rebar.

8. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcing bars according to claim 1, characterized in that, The construction of the obstacle domain in S1 includes: obtaining the dynamic positioning path of the component during the assembly process, scanning the reserved reinforcing bars of the lower structure in reverse along the positioning path, generating the reinforcing bar scanning body, and incorporating the reinforcing bar scanning body into the obstacle domain to ensure that the optimized reinforcement arrangement does not interfere with the assembly process.

9. A collision avoidance and multi-objective optimization device for spatial arrangement of reinforcing bars, characterized in that, include: Memory, used to store computer programs; The processor, coupled to the memory, is used to execute a computer program to implement the functions of the following modules: an environment construction module for constructing a parameterized reinforcement environment including a host domain, a barrier domain, and a protective layer boundary; and a population initialization module for generating an initial reinforcement population. The collision detection module performs multi-level collision detection and outputs the collision depth d and normal vector. The evaluation system module is used to establish multi-objective evaluation functions. The optimization solution module is used to run multi-objective optimization algorithms based on the collision depth d and the normal vector. Calculate the physical repulsive force vector and the repulsive force vector Superimposed on the position update equation of the individual steel reinforcement; The data output module is used to rationalize the optimized rebar geometry and output the result data; The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 8.

Images