Steel structure factory building design optimization method based on genetic algorithm
By integrating expert experience and hard constraints into a genetic algorithm, the design of steel structure workshops is optimized, solving the problems of randomness and data silos. This achieves efficient optimization of steel structure workshop design, improving design efficiency and compliance with standards.
Patent Information
- Application Number
- CN202511898284.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-16
- Publication Date
- 2026-03-20
AI Technical Summary
Traditional genetic algorithms suffer from randomness defects, experience gaps, and data silos in the design of steel structure workshops, resulting in low design efficiency and frequent need for manual adjustment of optimization results.
By integrating genetic algorithms with expert experience, the initial population is optimized by setting hard constraints and expert experience rules, and intervention is carried out during the crossover and mutation process. Combined with parametric modeling and finite element analysis, the specifications are met.
It improves the robustness of the design results and the efficiency of optimization, reduces the optimization cycle, and saves on engineering investment costs.
Smart Images
Figure CN121706198A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of civil engineering structure optimization, in particular to a steel structure plant design optimization method based on genetic algorithm and expert experience. BACKGROUND
[0002] Steel structure bent frame is the main structure form of shipyard plant building, which is characterized by large height span and load, resulting in large column and beam section size and wall thickness. The optimization design of the section of the main force components such as column and beam of the steel bent frame has an important influence on the investment cost of the shipyard. Design personnel design by computer modeling, and the optimization of the plant structure is usually to optimize the model by genetic algorithm, but the traditional genetic algorithm has three core defects in structure optimization: (1) Randomness defect: about 60%-70% of the individuals in the completely randomly generated initial population violate the basic engineering constraints (such as member slenderness ratio exceeding limit), resulting in waste of the first 20% of the iteration period in eliminating infeasible solutions. (2) Experience fault: although the existing improved scheme improves the uniformity of population distribution, it does not internalize the specification clauses such as GB50017 Steel Structure Design Standard, resulting in the need for manual secondary adjustment of the optimization results. (3) Data island: historical engineering data (such as typical node structure and economic span-height ratio) are not systematically utilized, resulting in loss of knowledge assets.
[0003] In view of this, in order to improve the reliability and practicability of genetic algorithm in practical engineering application, how to overcome the above shortcomings of genetic algorithm and improve the work efficiency of designers is a problem to be solved at present. SUMMARY
[0004] The present application provides a steel structure plant design optimization method based on genetic algorithm, which combines genetic algorithm and expert experience correction of steel structure plant design and is suitable for economic optimization of large-span industrial plants.
[0005] The technical problem to be solved by the present application is solved by the following technical scheme:
[0006] The present application provides a steel structure plant design optimization method based on genetic algorithm, which combines genetic algorithm and expert experience correction of steel structure plant design and is suitable for economic optimization of large-span industrial plants.
[0007] Step 1: Prepare plant classification data;
[0008] Step 2: Set expert experience rules and set hard constraint conditions;
[0009] Step 3: Construct an initial model by using a parameterized modeling method, and establish a finite element model based on the initial model;
[0010] Step four: adopt genetic algorithm to perform cyclic iteration optimization on the finite element model, and the optimization process meets the expert experience rules and hard constraint conditions;
[0011] Step five: visualize the optimization result.
[0012] Further, in the steel structure plant design optimization method based on the genetic algorithm provided by the application, the plant classification data in step one is divided into three types of light, medium and heavy according to the tonnage of the crane.
[0013] Further, in the steel structure plant design optimization method based on the genetic algorithm provided by the application, the expert experience rules in step two include giving recommended component sections for each type of plant according to historical engineering data.
[0014] Further, in the steel structure plant design optimization method based on the genetic algorithm provided by the application, the genetic algorithm optimization model in step four includes the following steps: initialization population, fitness evaluation, selection, crossover and mutation, in the initialization population step, the initial population is designed by using the expert experience rules, and the individuals that do not meet the conditions are removed in the initial population generation stage; in the crossover and mutation steps, the crossover and mutation results are intervened by using the expert experience rules.
[0015] Further, in the steel structure plant design optimization method based on the genetic algorithm provided by the application, the genetic algorithm optimization model process in step four meets the hard constraint conditions, and finally obtains the Pareto frontier.
[0016] Further, in the steel structure plant design optimization method based on the genetic algorithm provided by the application, the steel structure design standard specification clauses are used as the hard constraint conditions.
[0017] Further, in the steel structure plant design optimization method based on the genetic algorithm provided by the application, in step five, the optimization result is visualized in the form of Rhino three-dimensional.
[0018] The beneficial effects of the application are as follows:
[0019] The genetic algorithm-based steel structure plant design optimization method of the present application breaks the data island, formulates the initial population based on the expert experience rule, limits the section size of the steel structure component in the genetic algorithm calculation process and makes it meet the engineering requirements, secondly, takes the index of the specification as the hard constraint in the optimization process, and forces the optimization result to meet the specification requirements, through the above two aspects, the optimization method of the present application can significantly improve the robustness and optimization efficiency of the optimization result, thereby improving the work efficiency of the designers and achieving the purpose of saving the engineering investment cost. BRIEF DESCRIPTION OF DRAWINGS
[0020] Figure 1 is the genetic algorithm-based steel structure plant design optimization method flowchart in the embodiment of the present application;
[0021] Figure 2 is the display diagram of the expert experience rule and the hard constraint condition in the embodiment of the present application;
[0022] Figure 3 is the display diagram of the influence factor analysis of the optimization result in the embodiment of the present application;
[0023] Figure 4 is the visual display diagram of the optimization result in the embodiment of the present application. DETAILED DESCRIPTION
[0024] In order to make the technical means, creative features, purposes and effects realized by the present application easy to understand, the technical solutions of the present application are specifically described in the following embodiments in combination with the drawings.
[0025] Referring to Figure 1 A genetic algorithm-based steel structure plant design optimization method, comprising the following steps:
[0026] Step one: prepare the plant classification data:
[0027] According to the tonnage of the crane, the plant is divided into three types of light, medium and heavy.
[0028] Step two: set the expert experience rule and set the hard constraint condition:
[0029] The expert experience rule includes the expert experience rule used in the initialization of the population and the expert experience rule used in the cross variation process.
[0030] The expert experience rule for initializing the population: according to the historical engineering data, the recommended component section for each type of plant of light, medium and heavy is given. For example, when the tonnage of the crane reaches 100t~300t, the plant lower column section generally adopts the steel or steel pipe lattice column, the upper column generally adopts the solid web H-shaped steel section, the column section height and single limb section size are determined, and this component section generation process is embedded in the code for generating the initial population.
[0031] Expert experience rules in cross variation process: In the subsequent optimization process, the genetic algorithm cannot be left to cross variation at will, and intervention is still needed for the results of cross variation. For example, the cross-sectional size of the upper column should not be larger than the cross-sectional size of the lower column, the cross-sectional width of the steel beam should not be larger than the width of the upper column, and the results of cross variation can also be limited within a certain range or the maximum size of the cross section is limited. This also belongs to part of the expert experience rules.
[0032] Construct an expert experience rule library, and the following table shows some rules in the expert experience rules in this embodiment:
[0033]
[0034]
[0035]
[0036] The corresponding clauses of the Steel Structure Design Standard GB50017 and other standards (such as displacement limit 1 / 400, component local stability condition, slenderness ratio, etc.) are used as hard constraint conditions to force the optimization results to meet the standard requirements.
[0037] Referring to Figure 2 , the expert experience rule library (experience library) and the hard constraint condition (constraint relationship) are built into the optimization algorithm, Figure 2 Area A in the figure shows the case before the expert experience rule library and the hard constraint condition are built into the optimization algorithm, area B shows the case after the expert experience rule library and the hard constraint condition are built into the optimization algorithm, and area C is the design parameters of the experience library and the constraint relationship.
[0038] Step three: Construct an initial model, and establish a finite element model based on the initial model.
[0039] The parametric modeling method is used to construct the initial model, and the initial model is imported into the general finite element software for overall analysis and component design checking.
[0040] Step four: Use the genetic algorithm to perform cyclic iteration optimization on the finite element model, and the optimization process meets the expert experience rules and hard constraint conditions.
[0041] The genetic algorithm for model optimization includes the following steps: initialization of population, fitness evaluation, selection, cross, and variation.
[0042] In the initialization of the population step, the initial population is designed using the expert experience rules, and individuals that do not meet the conditions are removed in the initial population generation stage.
[0043] In the cross and variation steps, the expert experience rules are used to intervene in the cross variation results.
[0044] See Figure 3 During the iterative optimization process of the model, the final Pareto front is obtained under the premise of satisfying the hard constraints.
[0045] Step 5: Visualize the optimization results:
[0046] See Figure 4 The optimized design results of the steel structure factory building are visualized in three dimensions using Rhino (a 3D modeling software).
[0047] The above embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.
Claims
1. A design optimization method for steel structure factory buildings based on genetic algorithms, characterized in that, Includes the following steps: Step 1: Prepare factory classification data; Step 2: Set expert experience rules and set hard constraints; Step 3: Construct an initial model using parametric modeling methods, and then establish a finite element model based on the initial model; Step 4: Use a genetic algorithm to perform iterative optimization of the finite element model, ensuring that the optimization process satisfies expert experience rules and hard constraints; Step 5: Visualize the optimization results.
2. The steel structure factory building design optimization method based on genetic algorithm as described in claim 1, characterized in that: in, The factory classification data in Step 1 is based on the crane tonnage, dividing the factory into three types: light, medium, and heavy.
3. The steel structure factory building design optimization method based on genetic algorithm as described in claim 2, characterized in that: in, The expert experience rules in step two include providing recommended component sections for each type of factory building based on historical engineering data.
4. The steel structure factory building design optimization method based on genetic algorithm as described in claim 1, Its features are: Step four of the genetic algorithm optimization model includes the following steps: population initialization, fitness evaluation, selection, crossover, and mutation. In the initial population step, the initial population is designed using the expert experience rules, and individuals that do not meet the conditions are removed during the initial population generation stage. In the crossover and mutation steps, the expert experience rules are used to intervene in the crossover and mutation results.
5. The steel structure factory building design optimization method based on genetic algorithm as described in claim 4, characterized in that: in, In step four, during the genetic algorithm optimization process, the final Pareto front is obtained while satisfying the hard constraints.
6. The steel structure factory building design optimization method based on genetic algorithm as described in claim 1 or 5, characterized in that: in, The clauses of the steel structure design standards and specifications shall be used as the hard constraints.
7. The steel structure factory building design optimization method based on genetic algorithm as described in claim 1, characterized in that: in, In step five, the optimization results are visualized in Rhino 3D format.