A method for optimizing steel structure cutting paths based on genetic algorithms

By introducing concave corner bilateral imprint features and equipment lateral acceleration constraints into the genetic algorithm, the steel structure cutting path is optimized, solving the problems of insufficient expression of concave corner direction relationship and insufficient differentiation of acceleration requirements in the existing technology, and realizing more efficient cutting path planning.

CN123070447APending Publication Date: 2026-11-10JIANGSU NEW BLUE SKY STEEL STRUCTURE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611580551.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-13
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing steel structure cutting path optimization technology based on genetic algorithms cannot effectively express the directional relationship between the two sides of the concave corner, resulting in errors in the intergenerational transmission of surplus material features. Furthermore, it cannot distinguish the lateral acceleration requirements of the turning angle at different operating speeds, leading to a disconnect between the utilization of nesting features and equipment motion constraints.

Method used

By importing BIM models and equipment process parameters, a chromosome containing sequential coding segments and imprint coding segments is constructed. Concave double-sided imprint features are added to prioritize the generation of candidate poses. Combined with the lateral acceleration constraints allowed by the equipment, a genetic algorithm is used for iterative optimization to generate a cutting path scheme.

Benefits of technology

It improves the reusability of local geometric features of scrap material in next-generation layout, enhances the connection between path evaluation and equipment motion capability, and improves the compactness and smoothness of the cutting path.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123070447A_ABST
    Figure CN123070447A_ABST
Patent Text Reader

Abstract

This invention provides a method for optimizing steel structure cutting paths based on a genetic algorithm. The concave angles of the two individuals with the highest fitness in the previous generation are represented as bilateral imprint features containing the concave angle coordinates and the unit direction of the adjacent boundaries on both sides. During the next generation's decoding, candidate component poses matching these bilateral imprint features are preferentially generated. For the smoothed idle path, the curvature-constrained idle speed is determined based on the path curvature, the set idle speed, and the upper limit of the equipment's allowable lateral acceleration, and the idle lateral acceleration constraint index is calculated. The geometric features of the remaining material and the path constraint evaluation are fed back into the genetic iteration, improving the compactness of the layout and the smoothness of the idle path without adding dedicated sensors.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of steel structure cutting planning technology, specifically a method for optimizing steel structure cutting paths based on genetic algorithms. Background Technology

[0002] CNC cutting of steel structures is the core process of steel structure prefabrication. Material utilization, cutting efficiency and equipment operation stability directly affect processing costs and product quality. Existing steel structure cutting path optimization technology based on genetic algorithms usually uses BIM or CAD geometric data to establish component codes, and iteratively optimizes the layout or cutting sequence through selection, crossover and mutation. Individuals are evaluated by indicators such as layout occupancy range and idle travel. Existing technologies can also use polygon boundary points, concave corners or candidate points to assist in two-dimensional layout.

[0003] However, when only the coordinates of the feature points of the remaining material are saved, the directional relationship between the two sides of the concave corner cannot be expressed. The feature passed down across generations may provide incorrect guidance for the attitude of the next generation of components. At the same time, when evaluating the empty travel path only by the number of turning points or simple curvature, it is impossible to distinguish the difference in lateral acceleration requirements generated by the same geometric turning angle at different running speeds, resulting in a disconnect between the utilization of nesting features and the motion constraints of the equipment. Summary of the Invention

[0004] The purpose of this invention is to provide a steel structure cutting path optimization method based on genetic algorithm to solve the problem of insufficient expression of residual material imprint information.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for optimizing steel structure cutting paths based on genetic algorithms, comprising the following steps:

[0007] S1. Import the BIM model of the steel structure component to be cut, the original plate parameters and the cutting equipment process parameters, complete the coordinate system standardization process, and obtain the standardized component outline library and equipment process parameters.

[0008] S2. Construct chromosome coding rules based on a standardized component outline library and generate an initial population. The chromosome coding rules include sequential coding segments and imprinted coding segments. The imprinted coding segments of the initial generation are empty.

[0009] S3. Perform dynamic nesting decoding of leftover material component by component according to the chromosome cutting order; the actual initial leftover material area is always the original board area. When the imprint coding segment is not empty, read the leftover material concave corner bilateral imprint features of the two individuals with the highest fitness in the previous generation. Based on the concave corner point coordinates and the directions of the adjacent boundaries on both sides, candidate poses are generated first. Geometric inclusion, non-overlap verification, imprint adaptation evaluation and dynamic update of leftover material are completed to obtain the nesting arrangement result, the set of candidate points for concave corner of leftover material and the initial empty path node sequence.

[0010] S4. Identify turning points and smooth arcs in the initial empty travel path node sequence. Take the turning points that match the candidate points of the concave corner of the leftover material as the priority smoothing points. Determine the curvature constraint empty travel speed based on the curvature of the smoothed path, the set empty travel speed and the upper limit of the allowable lateral acceleration of the equipment. Calculate the empty travel lateral acceleration constraint index and generate the smoothed empty travel path parameters.

[0011] S5. Perform a two-layer comprehensive fitness evaluation on the nesting layout results and smoothed empty path parameters. The first layer removes individuals with components that cross the boundary, overlap, cannot be arranged, or whose path smoothing is not feasible. The second layer calculates the comprehensive fitness based on the nesting compactness rate, the normalized total arc length of the empty path, and the empty lateral acceleration constraint index to obtain the fitness ranking results of the population individuals.

[0012] S6. Genetic operator operations and imprint coding segment updates are performed based on the fitness ranking results to generate a new generation population. The lateral acceleration constraint index affects the source of the concave corner bilateral imprint feature of the next generation of leftover material through comprehensive fitness ranking. The new generation population is used as input to return to step S3 for iteration. After the termination condition is met, the cutting path scheme is output.

[0013] Further, in S1, the coordinates of the two-dimensional outer contour vertices and the projected area of ​​each component in the BIM model are extracted, and the original plate length and width, the cutting nozzle diameter of the cutting equipment, the minimum smooth arc radius set by the CNC system, the idle speed and the upper limit of the allowable lateral acceleration are imported; the equipment process parameters are read from the parameter table of the CNC system of the cutting equipment, the CAM process library or the equipment technical manual, and the component contour coordinates are converted into a two-dimensional rectangular coordinate system with the lower left corner of the plate as the origin.

[0014] Furthermore, in S2, each chromosome is composed of a sequential coding segment and an imprint coding segment concatenated together. The sequential coding segment is an integer arrangement of component numbers. The imprint coding segment stores the leftover concave corner bilateral imprint features of the two individuals with the highest fitness in the previous generation. Each bilateral imprint feature includes the coordinates of the concave corner point and the unit direction of the adjacent boundary segments on both sides of the concave corner. The unit direction of the adjacent boundary segments on both sides is taken as a unit vector pointing from the concave corner point to the corresponding adjacent vertex, and is numbered in ascending order of the direction angle of the unit vector relative to the positive direction of the horizontal axis of the two-dimensional rectangular coordinate system of the board as the first unit direction and the second unit direction.

[0015] Furthermore, the basic imprint adaptation in S3 The formula used is:

[0016]

[0017] Parallelism consistency in both directions is Imprint candidate poses are directly based on Sort.

[0018] Furthermore, the curvature constraint idle speed and idle lateral acceleration constraint indices in S4 are as follows: , Allowable upper limit of lateral acceleration It can be read from the parameter table of the CNC system of the cutting equipment, the CAM process library, or the equipment technical manual.

[0019] Furthermore, S5 employs a first-level feasibility threshold and a second-level comprehensive ranking; the second level is based on the layout compactness rate. Normalized total arc length of empty path Lateral acceleration constraint index during idle travel Calculate overall fitness The corresponding relationship is , , .

[0020] Furthermore, in S6, an elite retention and roulette wheel selection strategy is adopted to perform partial matching crossover and exchange mutation on the sequential coding segment; the leftover concave corner bilateral imprint feature set of the two individuals with the highest fitness in the current generation is updated as the imprint coding segment shared by all individuals in the next generation after deduplication.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. This invention adds the boundary direction information on both sides of the concave corner to the method of cross-generational transfer of feature points of leftover material, improves the point-level imprint to the concave corner bilateral imprint feature, and directly participates in the candidate pose priority determination with the parallel consistency of the bilateral direction. This makes the genetic information not only represent the position of the leftover material feature, but also the direction of boundary extension, thereby reducing the erroneous candidate poses generated by coordinate point matching alone, which is conducive to improving the reusability of the local geometric features of leftover material in the next generation of arrangement.

[0023] 2. This invention differs from schemes that evaluate paths solely based on the number of turns, idle travel length, or pure geometric curvature. It converts the set idle travel speed, path curvature, and the upper limit of the equipment's allowable lateral acceleration into an idle travel lateral acceleration constraint index, and incorporates this index into the genetic fitness evaluation. The same geometric path can yield different evaluation results under different set idle travel speeds or different allowable lateral acceleration conditions, directly linking path evaluation with the motion capability of the CNC equipment. Furthermore, it feeds back the geometric characteristics of the remaining material and the path constraint evaluation to the genetic iteration, improving the compactness of the nesting and the smoothness of the idle travel path. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0025] Figure 2 This is a schematic diagram of the process wireframe structure of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0027] Example:

[0028] Please see Figures 1-2 The present invention provides a technical solution:

[0029] A method for optimizing steel structure cutting paths based on genetic algorithms, comprising the following steps:

[0030] S1. Import the BIM model and processing resource parameters of the steel structure components to be cut, and obtain the standardized component outline library and equipment process parameters.

[0031] First, import the BIM model of the steel structure components to be cut, read the vertex coordinates of the two-dimensional outer contour polygons of each component, and the projected area of ​​the component to establish a component contour library. Then, import the original plate length and width, as well as the nozzle diameter of the cutting equipment. The minimum smooth arc radius set by the CNC system Set the idle speed and the upper limit of permissible lateral acceleration The above-mentioned equipment process parameters can be directly obtained from the equipment CNC system parameter table, CAM process library or equipment technical manual. Finally, the outline coordinates of all components are converted into a two-dimensional rectangular coordinate system with the lower left corner of the plate as the origin. The BIM model can contain or read the geometric outline and processing technology information of the components.

[0032] S2. Based on the standardized component outline library, construct chromosome coding rules containing sequential coding segments and imprinted coding segments to generate the initial population;

[0033] A genetic algorithm encoding vector is defined, which adds an imprint encoding segment to the cutting sequence encoding for transmitting local geometric information of the leftover material across generations. The initial generation imprint encoding segment is empty, so the initial population only relies on the sequence encoding and the conventional geometric arrangement rules to complete the first decoding.

[0034] Each chromosome consists of a sequential coding segment and an imprinted coding segment concatenated together. The sequential coding segment is an integer arrangement with a length equal to the total number of components to be cut. The value at each position corresponds to a component number, and the numbers are unique. The imprinted coding segment stores the bilateral imprinted features of the leftover concave corners of the two individuals with the highest fitness in the previous generation. For each leftover concave corner point, the coordinates of the concave corner point are recorded, and the two unit vectors pointing from the concave corner point to the two adjacent vertices on both sides are numbered in ascending order according to the direction angle relative to the positive direction of the horizontal axis of the two-dimensional rectangular coordinate system of the board, which is called the first unit direction and the second unit direction. This forms the bilateral imprinted features. The population size is twice the number of components and not less than 20. 40% of the individuals are initially ordered by the lowest horizontal line sorting algorithm, and the remaining 60% of the individuals are generated by random permutation. After completing the duplicate numbering check, new individuals are added to the set size.

[0035] S3. Dynamic nesting decoding of surplus material based on concave corner double-sided imprint matching;

[0036] The chromosome sequence encoding is converted into the actual nesting arrangement result. The bilateral imprint features are only used for priority guidance of candidate poses and are not used as the initial boundary of the actual surplus area to avoid the direct inheritance of the physical surplus state of the previous generation between different chromosomes.

[0037] At the start of decoding, regardless of whether the imprint encoding segment is empty, the actual initial surplus area is set to the complete area of ​​the original sheet material. For the component to be arranged, its allowable rotation angle set is read first. If the allowable rotation angle information is not read from the BIM processing attributes or CAM process file, only the allowable rotation angle is taken. When the imprint coding segment is not empty, the bilateral imprint features of each concave corner, each allowable rotation angle of the component to be arranged, and each component corner point are traversed. The component corner points are aligned with the concave corner points of the imprint, and direction matching is performed according to the aforementioned bilateral direction parallelism consistency to generate imprint candidate poses. The candidate poses are geometrically contained and non-overlapped using Vatti polygon Boolean intersection operation or equivalent polygon collision algorithm. After verification, the basic imprint fit degree and bilateral direction parallelism consistency degree are calculated, and the product of the two is selected from high to low. When the imprint coding segment is empty or all imprint candidate poses are infeasible, ordinary waste material boundary vertex candidate poses are generated. Ordinary waste material boundary vertex candidate poses refer to the candidate poses formed by traversing each allowable rotation angle of the component to be arranged and each component corner point, and aligning the component corner points with each boundary vertex of the current waste material region. If there are still no feasible poses in the ordinary waste material boundary vertex candidate poses, the chromosome is determined to be an infeasible individual.

[0038] Among them, basic imprint adaptation The calculation formula is as follows:

[0039]

[0040] To ensure the compatibility between the basic imprint of the components to be arranged and the current surplus material area. This is the current area of ​​remaining material. It is the shortest distance from the boundary point of the remaining material to the outline of the positioned components to be arranged. The diameter of the cutting nozzle of the cutting equipment. The perimeter of the remaining material area. This represents the current area of ​​surplus material. For the first Each cutting position corresponds to the projected area of ​​the component. The same component often has multiple candidate poses within the remaining material area. Each pose is scored to determine whether the component is close to the boundary of the remaining material and whether it fills the remaining material as much as possible. The larger the position, the better. The larger the value, the closer the outline of the located component is to the current boundary of the remaining material, and the higher the proportion of the component area to the current remaining material area. Indicates the degree of contribution of the mating edge, divided by Average the arc length to eliminate the influence of residual material size. Used to determine the degree of boundary fit; a value of 1 indicates that the entire boundary is fully fitted, and a value of 0 indicates that it is not fitted at all. Used to determine the area fill rate, if the area of ​​the current remaining material region is smaller than the projected area of ​​the component to be arranged, or if the candidate pose does not meet the requirements of geometric inclusion and non-overlap, the candidate pose is directly determined to be infeasible. For candidate poses that meet the conditions of geometric inclusion and non-overlap, ... The parallelism of the candidate imprint poses is further calculated in both directions. Its calculation formula is ,in, and They represent the first The first unit direction and the second unit direction of the concave angle of the imprint and They represent the first In the current rotation pose, the corner point of a component points in the first unit direction and the second unit direction pointing to the adjacent vertices on both sides. The two unit directions of the component corner point are also numbered from small to large according to the direction angle relative to the positive direction of the horizontal axis of the two-dimensional rectangular coordinate system of the plate. The parallelism of the bilateral directions is taken as 0 to 1. The closer to 1, the more parallel the two sets of boundary directions are.

[0041] After the component is placed, the Vatti 2D polygon Boolean clipping algorithm is used to subtract the outline of the placed component from the current surplus area and update the surplus area set. The updated surplus boundary is traversed, and interior angles greater than 100° are identified. For concave corner points, record the coordinates of the concave corner points and the unit directions pointing from the concave corner points to the adjacent vertices on both sides. Then, number the directions of the unit directions relative to the positive direction of the horizontal axis of the two-dimensional rectangular coordinate system of the board as the first unit direction and the second unit direction in ascending order to form a new candidate set of concave corner bilateral imprint features. After all components are arranged, extract the cutting retraction point of each component and the cutting entry point of the next component in sequence to generate an initial empty path node sequence.

[0042] S4. Analyze the circular arc smoothing and lateral acceleration constraints of the empty run based on the initial empty run path node sequence;

[0043] Path smoothing only applies to the idle travel section and does not change the cutting contour of the component itself. First, the turning points of the initial idle travel path are prioritized using the candidate points of the concave corner of the leftover material. Then, the curvature constraint idle travel speed of the smoothed path is determined based on the equipment motion parameters to avoid evaluating the lateral acceleration requirement of the path solely based on the size of the geometric turning angle.

[0044] Extract three consecutive nodes from the initial empty path node sequence, and calculate the discrete curvature of each intermediate node using the three-point circle method: ,in, Let the area of ​​the triangle formed by the three nodes be denoted as . , , When the three sides are collinear or have adjacent coincident points, the discrete curvature at that point is recorded as 0. The Euclidean distance between the turning point of the empty path and the candidate point of the concave corner of the remaining material is not greater than... When this happens, mark the inflection point as the priority smoothing point.

[0045]

[0046] When performing circular arc transitions, prioritize smoothing points first, then process the remaining non-collinear turning points. The radius of the inserted arc should not be less than [a certain value]. It requires that the arc be tangent to the preceding and following line segments. If the length of the current adjacent line segment is insufficient to form a radius not less than [a certain value], [further action is required]. If the transition arc is tangent to the preceding and following line segments, then the turning point is determined to not satisfy the arc smoothing constraint, and the corresponding individual is marked as path smoothing not feasible. After smoothing is completed, the line segment is taken as... , radius is Take the arc segment According to the set idle speed With the upper limit of permissible lateral acceleration The curvature constraint travel speed at each arc length position is determined using the above formula. When it is 0, Pick Next, let's analyze the lateral acceleration constraint parameters during idle travel: , The total arc length of the empty path. Let the coordinates be the arc length. For path curvature, the lateral acceleration constraint index for no-run is the average arc length of the lateral centripetal acceleration at each position on the path relative to the upper limit of the allowable lateral acceleration of the equipment. When it is 0, it is specified The value is 0. When the nozzle moves freely (without cutting, rapidly moving between components) through an arc or turn, it undergoes a curved motion with a lateral (centripetal) acceleration of 0. , This is used to quantify the average amount of acceleration capability used throughout the entire empty path; the smaller the value, the smoother the ride.

[0047] After the above processing is completed, update the total arc length of the smoothed idle path, the curvature at each arc length position, the curvature constraint idle speed, and the idle lateral acceleration constraint index, and use them as the fitness input for step S5.

[0048] S5. Based on the nesting layout results obtained in step S3 and the smoothed empty path parameters obtained in step S4, a two-layer comprehensive fitness evaluation is performed.

[0049] The first layer is the feasibility threshold: any individual with component out of bounds, component overlap, any component that cannot be arranged, or an empty path that cannot be completed smoothly according to step S4 is directly judged as infeasible. The second layer first calculates the layout compactness rate. ,in according to Calculate, then calculate the normalized total arc length of the empty path. .

[0050]

[0051] This represents the total number of components to be arranged. For the first The projected area of ​​each component , , and These represent the maximum and minimum x-coordinates, maximum and minimum y-coordinates, and minimum y-coordinates of all arranged component outline vertices in the plate coordinate system. and These are the original board length and width, respectively. The larger the value, the better the overall performance of the individual. The larger the value, the more compact the layout. The smaller the value, the shorter the empty path. The smaller the value, the lower the degree of lateral acceleration constraint occupancy.

[0052] S6. Perform genetic operator operations and imprinted coding segment updates based on fitness ranking results, generate a new generation population and return to S3 iteration. Output the cutting path scheme after the termination condition is met.

[0053] Population evolution is achieved through genetic operators, and the leftover material concave corner bilateral imprint features obtained from the real decoding of the two individuals with the highest fitness in the current generation are used to update the next generation imprint coding segment, so that the local geometric relationship of high-quality leftover material can be passed across generations. The lateral acceleration constraint index of idle walking changes the screening results of the two individuals through comprehensive fitness, thereby indirectly changing the source of the next generation imprint, forming a path constraint evaluation feedback on the subsequent nesting search direction.

[0054] The specific implementation process is as follows: First, a selection strategy combining elite retention and roulette wheel selection is adopted, retaining the top 10% of elite individuals, while the remaining individuals participate in roulette wheel selection according to their fitness ratio; then, partial matching crossover and exchange mutation are performed only on the sequential coding segment, with a mutation probability of 0.05. The imprinted coding segment does not participate in crossover and mutation. After completing the current generation evaluation, the bilateral imprint features of the two individuals with the highest fitness are merged, and the coordinate distance between them is not greater than 0.05 in the plate coordinate system. Furthermore, the concave angle feature, which is parallel to each other in the unit direction on both sides of the adjacent boundary, is used for deduplication, provided that the coordinate distance is not greater than [missing information]. However, concave corner features that do not satisfy the parallel relationship in the bilateral direction are retained, and bilateral direction information corresponding to individuals with higher fitness is retained to obtain the next generation shared imprint coding segment. When the absolute change of the packing compactness rate of three consecutive generations of adjacent best individuals is less than 0.001 and the absolute change of the lateral acceleration constraint index is less than 0.01, convergence is determined. If convergence is not achieved after 500 generations, the current best individual is output.

[0055] The final output includes the component nesting coordinate diagram corresponding to the optimal individual, the complete cutting path node sequence with arc transition, nesting compactness, actual material occupancy, idle travel distance, curvature constraint idle travel speed and idle travel lateral acceleration constraint index.

[0056] The component geometry data required for the above formula comes from BIM or CAD / CAM files. The plate size, nozzle diameter, set idle speed, minimum smooth arc radius, and allowable lateral acceleration upper limit come from the equipment CNC parameters, CAM process library, or equipment technical manual. The excess material boundary, concave corner double-sided imprint features, path curvature, and envelope rectangle area are all calculated from the aforementioned geometry data. Before calculation, the length, speed, and acceleration are converted according to the same unit system.

[0057] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0058] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0059] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for optimizing steel structure cutting paths based on genetic algorithms, characterized in that, The specific steps include: S1. Import the component BIM model, original plate parameters and equipment process parameters, and obtain the component outline library and equipment process parameters through coordinate standardization. S2. Construct chromosome coding rules containing sequential coding segments and imprinted coding segments based on the component contour library, and generate an initial population. The initial generation imprinted coding segments are empty. S3. Dynamic nesting decoding of components according to the cutting order, the initial residual material is always the original plate; when the imprint segment is not empty, read the concave corner double-sided imprint of the two best bodies of the previous generation, generate candidate poses according to the concave corner coordinates and the two sides, update the residual material after inclusion, non-overlap verification and adaptation evaluation, and obtain the nesting layout result, the concave corner candidate point set of residual material and the initial empty path node sequence. S4. Identify turning points in the initial empty running node sequence and smooth them with arcs. Take the turning points that match the candidate points of the concave corner of the leftover material as the priority smoothing points. Determine the curvature constraint empty running speed based on the curvature of the smoothed path, the set empty running speed and the upper limit of the allowable lateral acceleration of the equipment. Calculate the empty running lateral acceleration constraint index and generate the smoothed empty running path parameters. S5. Perform a two-layer fitness evaluation on the nesting layout results and the smoothed empty path parameters to obtain the fitness ranking of individual populations. S6. Based on fitness ranking, execute genetic operators and update imprint coding segments to generate a new generation population; the lateral acceleration constraint index of idle walking affects the source of the concave corner bilateral imprint feature of the next generation of leftover material through comprehensive fitness ranking, and returns the new generation population as input to step S3 iteration. After the termination condition is met, the cutting path scheme is output.

2. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 1, characterized in that: In step S1, the coordinates of the two-dimensional outer contour vertices and the projected area of ​​each component in the BIM model are extracted. The original plate length and width, the cutting nozzle diameter of the cutting equipment, the minimum smooth arc radius set by the CNC system, the idle speed and the upper limit of the allowed lateral acceleration are imported. The component contour coordinates are then converted into a two-dimensional rectangular coordinate system with the lower left corner of the plate as the origin.

3. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 1, characterized in that: In S2, each chromosome is composed of sequential coding segments and imprinted coding segments concatenated together. The sequential coding segments are integer arrangements of component numbers. The imprinted coding segments store the leftover concave corner bilateral imprint features of the two individuals with the highest fitness in the previous generation. Each bilateral imprint feature includes the coordinates of the concave corner point and the unit direction of the adjacent boundary segments on both sides of the concave corner. The unit direction of the adjacent boundary segments on both sides is taken as the unit vector pointing from the concave corner point to the corresponding adjacent vertex, and is numbered in ascending order of the direction angle of the unit vector relative to the positive direction of the horizontal axis of the two-dimensional rectangular coordinate system of the board as the first unit direction and the second unit direction.

4. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 3, characterized in that: In S3, the basic imprint fit of the candidate pose for: To ensure the compatibility between the basic imprint of the components to be arranged and the current surplus material area. This is the current area of ​​remaining material. It is the shortest distance from the boundary point of the remaining material to the outline of the positioned components to be arranged. The diameter of the cutting nozzle of the cutting equipment. The perimeter of the remaining material area. This represents the current area of ​​surplus material. For the first The projected area of ​​the component corresponding to each cut position is used for candidate poses that satisfy the conditions of geometric inclusion and no overlap. , for the first The first bilateral imprint feature and the first Candidate poses generated from the corner points of each component are used to calculate the parallelism in both directions. Its calculation formula is ,in, and The first The first unit direction and the second unit direction of the concave angle of the imprint and The first In the current rotational pose, each component corner point points to its first and second unit directions from the corner point to its two adjacent vertices. The unit directions on both sides of the component corner point are also numbered in ascending order of their direction angles relative to the positive x-axis of the two-dimensional Cartesian coordinate system of the board. Priority is given to candidate imprint poses that satisfy geometric inclusion and non-overlapping conditions. If no feasible imprint candidate pose is found, then the candidate poses of ordinary residual material boundary vertices are used for arrangement.

5. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 4, characterized in that: In step S4, the set idle speed of the cutting equipment is used. and the upper limit of permissible lateral acceleration Determine the curvature constraint for the idle travel speed of the smoothed path, when the path curvature When the value is not zero, the curvature constraint idle speed is: when When it is zero, Pick Calculation of lateral acceleration constraint index for air travel based on curvature constraint air travel speed : ,in, The total arc length of the empty path. Let the coordinates be the arc length. For path curvature, when Zero time regulation When it is zero, When greater than zero , This is used to represent the average arc length occupied by the smoothed idle path under curvature-constrained idle speed, which affects the allowable lateral acceleration of the equipment.

6. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 5, characterized in that: In S5, the compactness of the layout Area of ​​the envelope rectangle Normalized total arc length of empty path and overall adaptability According to , , , , Calculate, where, This represents the total number of components to be arranged. For the first The projected area of ​​each component , , and These represent the maximum and minimum x-coordinates, maximum and minimum y-coordinates, and minimum y-coordinates of all arranged component outline vertices in the plate coordinate system. and These are the original board length and width, respectively. The larger the value, the better the overall performance of the individual.

7. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 6, characterized in that: In step S6, a selection strategy combining elite retention and roulette wheel selection is employed. Partial matching crossover and exchange mutation are performed on the sequential coding segment. The residual material concave corner bilateral imprint features of the two individuals with the highest fitness in the current generation are merged. In the plate coordinate system, the coordinate distance is not greater than... Furthermore, concave angle features that are parallel in unit direction to adjacent boundaries on both sides are considered duplicate features and are removed. For coordinate distances not greater than [missing information], [missing information]. However, concave corner features that do not satisfy the parallel relationship in the bilateral direction are retained respectively, and bilateral direction information corresponding to individuals with higher fitness is retained first. The resulting feature set is used as the imprint coding segment shared by all individuals in the next generation.

8. The method for optimizing steel structure cutting paths based on genetic algorithms according to claim 1, characterized in that: The two-layer fitness evaluation includes a first layer and a second layer. The first layer removes individuals that are out of bounds, overlap, cannot be arranged, or are not smooth and can not be executed. The second layer calculates the comprehensive fitness based on the compactness rate, the normalized total arc length of the idle path, and the idle lateral acceleration constraint index.