Search

CN-121997431-A - Collision avoidance and multi-objective optimization method and device for reinforcement space arrangement

CN121997431ACN 121997431 ACN121997431 ACN 121997431ACN-121997431-A

Abstract

The invention discloses a collision avoidance and multi-objective optimization method and device for reinforcement space arrangement, and relates to the technical field of digital design and intelligent construction of building structures. The method comprises the steps of firstly constructing a parameterized three-dimensional reinforcement distribution environment comprising a concrete host domain, an obstacle domain and a protective layer, adopting a multistage collision detection strategy based on spatial indexes, quickly locking an interference object through rough screening of a bounding box and geometric accurate intersection, establishing a multi-objective evaluation system integrating standard hard constraint and construction accessibility soft constraint, converting a collision detection result into a physical repulsive force to directly drive reinforcement position update by utilizing an improved multi-objective optimization algorithm introducing a physical collision feedback mechanism, and finally generating numerical control machining data conforming to BVBS standard. According to the invention, the algorithm convergence efficiency is obviously improved through a physical feedback mechanism, and the automatic collision-free arrangement of the reinforcing steel bars meeting the construction and processing requirements under a complex special-shaped structure is realized.

Inventors

  • XU JIN
  • WEI JIANJUN
  • HE XU
  • LI XIONGWEI
  • QIAN XIAO

Assignees

  • 常州工程职业技术学院

Dates

Publication Date
20260508
Application Date
20260128

Claims (10)

  1. 1. The collision avoidance and multi-objective optimization method for the spatial arrangement of the reinforcing steel bars is characterized by comprising the following steps of: s1, constructing a parameterized reinforcement distribution environment, namely acquiring geometric information of a concrete member, determining a host domain in which the reinforcement is arranged, identifying an embedded part and a hole to form an obstacle domain, and inwards shifting the boundary of the host domain according to the thickness of a protective layer required by design specifications to form a three-dimensional effective space allowed to exist by a reinforcement center line; s2, generating an initial reinforcing steel bar population in an effective space, wherein each reinforcing steel bar individual is stored in a parameterized form and comprises a space position coordinate, an axial direction vector and a section geometric parameter; S3, performing multistage collision detection, namely performing collision detection on the reinforcing steel bar population, performing coarse screening by adopting bounding box space indexes, performing geometric precise intersection operation on potential interference objects, and outputting collision state information of each reinforcing steel bar, wherein the collision state information comprises a collision depth scalar d and a normal unit vector at a collision point ; S4, establishing a multi-objective evaluation system, namely constructing an evaluation function comprising mandatory engineering constraint and non-mandatory optimization objective, wherein the mandatory engineering constraint is used as a feasibility criterion of a solution, and the non-mandatory optimization objective is used as an adaptability calculation basis; S5, performing iterative optimization introducing physical collision feedback, namely running a multi-objective optimization algorithm to perform iterative solution, wherein in each iterative process, the collision depth d and the normal unit vector acquired in the S3 are used for performing iterative solution Calculating a physical repulsive force vector Vector of physical repulsive force Vector superposition is carried out on the initial position updating component of the optimization algorithm, so that the actual position updating quantity of the individual reinforcing steel bars is formed, and the reinforcing steel bars are evolved to a collision-free area on the premise of meeting an optional optimization target until a preset convergence condition is met; and S6, outputting a reinforcement arrangement result, namely rationalizing the reinforcement geometry after optimization convergence, and outputting reinforcement arrangement result data.
  2. 2. The collision avoidance and multi-objective optimization method for rebar spatial arrangement according to claim 1, wherein the physical repulsive force vector in S5 The calculation formula of (2) is as follows: ; Wherein: As the repulsive force weight coefficient, the repulsive force weight coefficient is decreased with the increase of the iteration times; d is a collision depth scalar; normal unit vector pointing to the repelled steel bar at the collision point, and fixed the hardness coefficient of the barrier Is larger than the hardness coefficient among the individual steel bars; Wherein the repulsive force weighting coefficient The iteration times are followed by adopting any one of the following modes Decreasing: (1) Linear attenuation: ; in the formula, As a result of the initial weight coefficient, The maximum iteration number; (2) Exponential decay: ; in the formula, For the attenuation rate coefficient, the value range is ; Hardness coefficient of collision object The value rule of (2) is as follows: (1) When the collision object is a concrete boundary or obstacle domain, , The value range is ; (2) When the collision object is an individual of other reinforcing bars, , The value range is And meet the following 。
  3. 3. The collision avoidance and multi-objective optimization method for rebar spatial arrangement according to claim 1, wherein the multi-stage collision detection in S3 comprises: The first stage of spatial index coarse screening, namely constructing an axial bounding box for each steel bar, performing spatial division by utilizing an octree or hierarchical bounding box BVH structure, judging that steel bar pairs which are unlikely to collide in space are rapidly removed through bounding box intersection, and outputting a potential interference object set; The second stage of geometric accurate intersection, namely performing accurate geometric intersection operation on the steel bar pairs in the potential interference object set, selecting a line segment-line segment, line segment-plane or cylinder-cylinder intersection algorithm according to the geometric form of the steel bar, and calculating the collision depth Normal unit vector at collision point 。
  4. 4. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcement bars according to claim 1, wherein when the concrete member is a special-shaped curved member, the generation method of the initial reinforcement bar population in S2 is specifically that a streamline field is generated by extracting the principal curvature direction of the surface of the host domain or according to the principal stress distribution of the structure, and the spatial topology structure and trend of the individual reinforcement bars are initialized along the streamline field direction, so that the initial reinforcement bar arrangement conforms to the curved shape of the member.
  5. 5. The collision avoidance and multi-objective optimization method for rebar spatial arrangement according to claim 1, wherein the mandatory engineering constraints in S4 comprise: The thickness constraint of the protective layer is that the distance from the center line of the steel bar to the boundary of the concrete is not smaller than the thickness of the designed protective layer, the clear distance constraint of the steel bar is that the minimum clear distance between the surfaces of adjacent steel bars is not smaller than a standard specified value, the anchoring length constraint is that the anchoring length of the end part of the steel bar meets the structural design requirement, and the safety distance constraint of the barrier is that the minimum distance between the steel bar and the boundary of the barrier domain is not smaller than a preset safety allowance; The non-mandatory optimization target at least comprises one of construction accessibility, machining complexity, material consumption and evaluation index, wherein the evaluation index is minimized by the interference volume of a virtual operation sphere and a reinforcing steel bar at a key node, the evaluation index is minimized by the occurrence frequency of a non-standard bending angle and the total bending frequency of the reinforcing steel bar, and the evaluation index is minimized by the total length of the reinforcing steel bar.
  6. 6. The collision avoidance and multi-objective optimization method for rebar spatial arrangement according to claim 1, wherein the specific method in S6 comprises: (1) Geometry rationalization, namely fitting an optimized steel bar center line curve into a combination of a straight line segment and a circular arc segment, correcting a bending angle, forcedly correcting the bending angle into a standard angle when the deviation between the bending angle and 90 degrees or 135 degrees is smaller than a preset angle tolerance, and constraining the minimum bending radius according to the diameter of the steel bar; (2) And outputting data, namely analyzing the geometric parameters, the material information and the bending sequence of the steel bar into an instruction format which accords with the standard of numerical control machining equipment, wherein the instruction format comprises BVBS format or IFC format.
  7. 7. The collision avoidance and multi-objective optimization method for spatial arrangement of reinforcement according to claim 1, wherein in the iterative optimization process in S5, when it is detected that the collision depth of a certain reinforcement is not significantly reduced continuously for N generations, a topology variation operation is performed on the reinforcement, the topology variation operation including inserting a control point on a reinforcement centerline or adjusting a reinforcement end anchoring manner to increase the geometrical freedom of the reinforcement.
  8. 8. The collision avoidance and multi-objective optimization method for rebar spatial arrangement according to claim 1, wherein the constructing of the obstacle domain in S1 comprises obtaining a dynamic seating path of the component during assembly, reversely scanning reserved tie bars of the underlying structure along the seating path, generating a tie bar scanning body, and incorporating the tie bar scanning body into the obstacle domain to ensure that the optimized rebar arrangement does not interfere with the assembly process.
  9. 9. Collision avoidance and multi-objective optimization device for reinforcement space arrangement, characterized by comprising: A memory for storing a computer program; A processor coupled with the memory for executing a computer program to realize the functions of an environment construction module for constructing a parameterized reinforcement distribution environment including a host domain, an obstacle domain, and a protective layer boundary, a population initialization module for generating an initial reinforcement population, a collision detection module for performing multistage collision detection and outputting a collision depth d and a normal vector The evaluation system module is used for establishing a multi-objective evaluation function; the optimization solving module is used for running a multi-objective optimization algorithm according to the collision depth d and the normal vector Calculating a physical repulsive force vector And vector the repulsive force The data output module is used for rationalizing the optimized steel bar geometry and outputting result data; Wherein the processor, when executing the computer program, implements the steps of the method of any one of claims 1 to 8.
  10. 10. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 8.

Description

Collision avoidance and multi-objective optimization method and device for reinforcement space arrangement Technical Field The invention relates to the technical field of intelligent construction of civil engineering, in particular to a collision avoidance and multi-objective optimization method and device for reinforcement space arrangement. Background With the acceleration of the building industrialization process, the arrangement work of the steel bars is gradually transferred from the two-dimensional plain-law diagram to the three-dimensional digital model. It has become a common practice to conduct collision checks with BIM software. For example, patent number CN114329693B realizes collision recognition in the process of steel bar sample-turning by performing boolean operation on the solid model, and patent number CN113032861a proposes a component collision analysis method based on BIM model. The technology can effectively identify interference conditions among the reinforcing steel bars and the embedded parts, but is usually limited to output collision positions, and subsequent avoidance adjustment still depends on manual completion of designers. In the case of complex nodes or variable component shapes, manual correction is labor intensive and local modifications may lead to new interventions. On the other hand, an automatic reinforcement distribution idea based on a rule base also exists in the industry. For example, patent number CN104156544a automatically generates beam column node rebar through a preset three-dimensional node module. The method has certain efficiency in the rule component, but because the rule base is limited, once a special-shaped curved surface, a non-standard intersecting node or a scene with various obstacles is encountered, the original rule is difficult to cover the actual requirement, and manual intervention and trimming are still needed. In addition, some researches attempt to improve the automation level of the reinforcement distribution by using a meta heuristic algorithm, such as adopting an improved genetic algorithm to perform structural optimization design in patent number CN114611191a, and related academic works also explore to solve the problem of reinforcement congestion by using a particle swarm algorithm. The algorithm can process multi-objective optimization to a certain extent, but the iterative process lacks direct perception on the geometric environment, generally relies on random search promotion, so that the convergence speed is low, local optimization is easy to fall into, and the formed steel bar form sometimes is difficult to meet the requirements of processability or construction accessibility. Under the technical conditions, the deep design of the steel bars still faces obvious bottlenecks in complex components. For a high-density reinforcement region, a traditional optimization algorithm is difficult to find an arrangement scheme which is free of collision and meets the standard requirement in a limited time, and the situation that calculation is stagnant or repeated adjustment is easy to occur in an iteration process, so that interference cannot be thoroughly eliminated is often caused. As the component geometry becomes more complex, solutions that rely purely on penalty functions or random searches have difficulty in effectively driving the solution to converge towards a feasible region. Even if a geometrically "collision-free" arrangement is possible, the existing methods lack effective guarantees in terms of workability. Many optimization results only pay attention to the avoidance relation in space, but cannot synchronously consider construction factors such as the processing radius of the steel bar, the occurrence frequency of nonstandard bending, the node operation space and the like. In actual construction, too many bending angles or insufficient operation surfaces can lead to the fact that the scheme cannot fall to the ground, so that reworking is generated in the construction stage. Meanwhile, the problem of data disconnection still exists between the design model and the processing equipment in the current flow. The rebar placement results often need to be manually rearranged into a format for numerical control processing, and such manual conversion in complex components is time consuming and prone to errors. For the integrated process of pursuing digital construction, how to make the arrangement result directly correspond to the standardized processing instruction is also an unresolved difficulty in the prior art. In summary, in the prior art, when the complex space conditions, high-density reinforcement and components with multiple types of obstacles coexist, the problems of insufficient automation degree, low anti-collision solving efficiency, difficulty in considering structural specifications and construction processes and the like still exist. How to effectively obtain geometrical feedback such as collision position, collision depth