Beam plate type steel structure optimization design method based on informatization modeling and parameterization representation

Abstract

The invention discloses a beam plate type steel structure optimization design method based on informatization modeling and parameterization representation. The invention provides a hierarchical modeling and parameter mapping mechanism based on a graph data structure. Components such as beams, plates and columns are abstracted into vertexes, edges and directed closed loops, space coordinates, linear parameters, surface thicknesses and material attributes of the components are defined respectively, and a graph model containing a component attribute information set is constructed. A Tarjan strong connected component algorithm is adopted to automatically search and verify a closed loop, and the physical rationality of a structural surface is ensured. Design parameters are dynamically associated with vertex coordinates, side line shapes and closed-loop thickness through a mapping function, and parameter-driven geometric shape adjustment and batch model generation are supported. According to the method, efficient modeling, comprehensive information representation and real-time dynamic updating of a complex structure are achieved, the design precision and the cooperation efficiency are remarkably improved, and the method can be widely applied to design optimization and construction management of large beam-slab steel structures in ocean engineering, building engineering and traffic engineering.

Description

Technical Field

[0001] The present invention relates to the fields of building information modeling (BIM) technology and steel structure design, and in particular to an optimization design method for beam-slab steel structures based on information modeling and parametric characterization. Background Art

[0002] Beam-slab steel structures are a common building structure composed of beams and slabs, primarily used to bear loads and transfer them to supporting structures. Their advantages include high strength, light weight, fast construction, and excellent seismic resistance, making them widely used in a variety of fields, including marine engineering, construction, and transportation engineering. In marine engineering, beam-slab steel structures are commonly used in facilities such as offshore oil platforms and port terminals. For example, offshore oil platforms must withstand the complex loads of the marine environment, and beam-slab steel structures provide sufficient strength and stability. In construction engineering, beam-slab steel structures are widely used in high-rise buildings, large stadiums, and industrial plants. For example, large stadiums often require long-span roof structures, and beam-slab steel structures can meet these design requirements. In transportation engineering, beam-slab steel structures are used in infrastructure such as bridges and highways, including long-span bridges and elevated roads.

[0003] With the continuous expansion of project scale and increasing structural complexity, traditional design and modeling methods, especially for large beam-slab steel structures, are no longer able to meet the requirements for efficient and accurate engineering. Large structures often involve a large number of components and complex connections, significantly increasing the need for optimized design. Furthermore, complex structural design and construction requirements place higher demands on modeling and representation methods. Currently, traditional two-dimensional design and simple three-dimensional modeling methods have many shortcomings when dealing with large and complex structures. For example, two-dimensional drawings struggle to intuitively represent complex three-dimensional spatial relationships, information transmission is prone to errors, and rapid modification and updating are difficult. While existing three-dimensional modeling methods have improved design efficiency to a certain extent, they still lack efficient parametric representation and information management capabilities, making them unable to meet the requirements for refined design and optimization of complex structures. Furthermore, parametric representation enables rapid batch model generation and modification, providing strong support for structural optimization design and construction management. This approach can effectively address the shortcomings of traditional methods, meet the design and construction requirements of modern engineering for large and complex structures, and has important practical application value. Summary of the Invention

[0004] The present invention provides a beam-slab steel structure optimization design method based on information modeling and parametric characterization. By establishing a graph data structure, adopting a hierarchical information integration strategy and developing a model parameter mapping characterization mechanism, efficient modeling, comprehensive information characterization and dynamic updating of beam-slab steel structures are achieved, meeting the optimization needs of large-scale structures, improving design efficiency and quality, and enhancing the synergy between design and construction.

[0005] The technical solutions adopted by the present invention to solve the technical problems are as follows:

[0006] A beam-slab steel structure optimization design method based on information modeling and parametric characterization includes:

[0007] (1) Based on the graph data structure, the information model of the beam-slab steel structure is obtained;

[0008] (2) mapping the design parameters to the structural information model obtained in step (1) to obtain a structural information basic model containing the initial design parameters;

[0009] (3) Based on the structural information basic model containing the initial design parameters obtained in step (1), the design parameters are adjusted and optimized through an automatic optimization algorithm to finally obtain an optimized structural information model. The beam-slab steel structure is optimized according to the optimized structural information model.

[0010] The present invention proposes a hierarchical modeling and parameter mapping mechanism based on a graph data structure. By abstracting components such as beams, plates, and columns into vertices, edges, and directed closed loops, their spatial coordinates, linear parameters, surface thickness, and material properties are defined respectively, and a graph model containing a component attribute information set is constructed. The Tarjan strongly connected component algorithm is used to automatically search and verify closed loops to ensure the physical rationality of the structural surface. The design parameters are dynamically associated with the vertex coordinates, edge shape, and closed loop thickness through a mapping function, supporting parameter-driven geometric shape adjustment and batch model generation. This method achieves efficient modeling of complex structures, comprehensive information representation, and real-time dynamic updating, significantly improving design accuracy and collaborative efficiency, and can be widely used in the design optimization and construction management of large-scale beam-slab steel structures in marine engineering, construction engineering, and transportation engineering.

[0011] The information modeling of beam-slab steel structures based on graph data structures abstracts the various components (such as beams, plates, columns, connectors, etc.) in the beam-slab steel structure into structural nodes (points), structural line segments (line segments), and structural surfaces (surfaces). Vertices (vertices), edges (edges), and directed closed loops (directed closed loops) in the graph data structure correspond to the nodes, line segments, and surfaces of the structure, respectively, and store component information at different locations. The specific steps include:

[0012] (1) Use vertices to perform information modeling on the structural nodes, and the expression is:

[0013]

[0014] Where, v iis any vertex in the space, V is the set of all vertices in the structure, a total of n v vertices. x i ,y i ,z i is the spatial coordinate of this vertex. For vertex v i As a component attribute information set when connecting nodes.

[0015] (2) Use the edge to perform information modeling on the structural line segment, which can be expressed as:

[0016]

[0017] Where, v i and v j To form edge e ij The two vertices of . E is the set of all edges in the structure. is a set of structural surface labels, indicating e ij One or more tags are located at ET1 to On the surface, Is a character variable; ES represents the line type of the edge, and integers ES = 1, 2, 3, 4 represent straight line segments, circular arcs, elliptical arcs, and Bezier curves respectively; P ij ={p1,p2,...} are parameters that control the shape of a specific type of curve and are used to define circular arcs, elliptical arcs, and Bezier curves. For edge e ij A set of component property information when used as a structural component (beam, column, etc.).

[0018] (3) Define the directed closed loop in the graph data structure. The expression of the directed closed loop is:

[0019]

[0020] Where c k Represents the kth directed closed loop in the graph, consisting of a series of edges Composition, subscript i i ...i n Refers to the vertices of the edge, where the starting and ending vertices are the same. In addition to the defined edges, the loop has two other structural construction properties: CE k is the surface type, where the integer CE k =1,2 represent plane surface and curved surface respectively; is a directed closed loop c k A set of component attribute information when representing a structural surface.

[0021] (4) Automatic search for directed closed loops in graph data structures and their verification:

[0022] For the graph data structure constructed by steps (1), (2) and (3), the Tarjan strongly connected components (SCCs) algorithm is used to automatically search for directed closed loops in the graph data structure, aiming to find all strongly connected components of the directed graph through the depth-first search (DFS) method. For each directed closed loop found, the intersection of the thickness sets of all edges on the loop is checked twice. If the calculated intersection is not an empty set, it can be considered that all edges of this directed closed loop share a thickness label, which means that this loop can form a surface. The expression of the secondary check is written as:

[0023]

[0024] In the formula, the subscript of ET represents the vertex number of the corresponding edge. Since the corresponding edges form a closed loop, the first vertex number of the first edge is the same as the vertex number of the last edge, i.e. i1. The loop that does not meet the above check will not be included in the depth-first search stack, while the loop that passes the check will be assigned a component attribute information set

[0025] The definition form of the component attribute information set in the above steps (1), (2) and (3) specifically includes:

[0026] (1) Vertex v i Component attribute information set Its expression is:

[0027]

[0028] Where, Identify the variable for the connection type, Identify unconnected free vertices, Only identification variables are included in They represent welding, bolt connection and bolt-weld connection respectively. It is the welding related information parameter, It is the information parameter related to the bolt connection. A set of material properties for the connection, defined only when the vertices represent connection nodes.

[0029] (2) Edge e ij Component attribute information set Its expression is:

[0030]

[0031] Where, Identify the variable for the beam's cross-section type, Represents edge e ij Not used as a structural member; Indicates standard cross-section form. They represent standard sections such as I-section, H-section, C-section, and circular tube section respectively. The related parameter sec is the model of the standard section and is a string format variable. Indicates a non-standard section, defined by a polyline. x,k and p y,k Indicates the local coordinates of the end points of the cross-section polygon, t k is the thickness at the end point of the kth section, and the section definition has n edge endpoints. A set of material properties for edges, defined only when the edges represent structural members.

[0032] (3) Directed closed loop c k Component attribute information set Its expression is:

[0033]

[0034] Where, t k Indicates the thickness of the panel. c in is the set of internal closed loops, representing the inner boundary of the panel. in Can be an empty set, meaning the panel has no internal boundaries. If the internal boundaries form a partial hollow, set the thickness of the internal closed loop to 0. is a material property set of a directed closed loop. When the closed loop is hollowed out,

[0035] Design parameter mapping characterization method, specifically including:

[0036] (1) Forming structural geometric design parameters Directly with the vertex v i The spatial coordinate x i ,y i ,z i Mapping relationship. Determine the design parameter set based on the design drawing Including length, angle, curvature, etc., establish a mapping function or model to associate these geometric design parameters with the spatial coordinates of the vertices. By adjusting the design parameters, the spatial coordinates of the vertices will change accordingly, thereby achieving the adjustment and optimization of the overall geometric shape of the structure. Mapping function f i The overall expression of can be written as:

[0037]

[0038] (2) Determine linear ES based on design drawings ij , distinguish between straight line segments, circular arcs, elliptical arcs and arbitrary Bezier curves, and according to ES ij Determine the control parameter set P of the corresponding edge ij For a straight line segment, ESij =1,P ij is an empty set; for arcs, that is, ES ij =2,P ij Contains only the coordinates of the circle center; for elliptical arcs, ES ij =3,P ij Contains the coordinates of the center point and the semi-major axis endpoints; for Bezier curves, that is, ES ij =4,P ij Contains the number of control points, control point coordinates, and curve control parameters, defining the edge according to the shape of the structural member ij The linear parameter ES in ij and P ij ;

[0039] (3) Determine the surface type CE according to the design drawings k And the thickness set t of the panel boundary nodes under the corresponding parameter value k For example, a quadrilateral with uniform thickness t0, t k =(t0,t0,t0,t0); if a pentagon has different thicknesses at each end, the thicknesses of the end points are t1, t2, t3, t4 and t5 respectively, then t k =(t1, t2, t3, t4, t5). Determine whether there is a loop inside the face to construct the internal closed loop set c in .

[0040] (4) Establish a standard library of nodes and beam sections, assign standard numbers, build a standard library of steel structure materials, and set the material properties of structural nodes, segments and surfaces. and Select and define from the standard material library, and determine the attribute information set based on the cross-section drawing of the design drawing and

[0041] Based on the structural informationization basic model containing the initial design parameters, the design parameters are adjusted and optimized using an automatic optimization algorithm to ultimately obtain an optimized structural informationization model. The beam-slab steel structure is then optimized based on the optimized structural informationization model. The specific steps include:

[0042] (1) Establish an optimization problem. Single-objective or multi-objective optimization problems can be constructed based on objectives such as minimizing structural mass and optimizing structural mechanical response, with design specification requirements and actual construction conditions serving as constraints. In single-objective optimization, constraints are directly integrated with the optimization objectives in the form of penalty functions to ensure that the optimization results meet the specifications and working conditions. In multi-objective optimization, constraints also guide the optimization algorithm to find the optimal equilibrium solution while satisfying all constraints by weighing the priorities of different objectives.

[0043] (2) Determine the parameter selection range based on the initial design drawings. For any parameter selection combination, the parameter mapping method is used to build an information model, and the geometry and component information are read based on the 3D modeling software script data interface to perform geometric modeling and divide the finite element mesh.

[0044] (3) Solve the fitness function of the optimization problem. Use the finite element mesh obtained in step (2), apply boundary conditions and design loads, and conduct finite element simulation. Read the key structural response indicators based on the finite element simulation results. Automatically calculate the overall quality of the structure based on the information model of the structure. Construct a fitness function based on the structural quality and key structural response indicators according to the optimization problem description method described in step (1).

[0045] (4) Update the design parameters based on the optimization algorithm. Use a discrete value swarm optimization algorithm, such as a genetic algorithm, ant colony algorithm, and particle swarm algorithm, to perform steps (1) to (3) on different individuals. Integrate the fitness function solution results of all individuals and update all design parameters according to the parameter update strategy of the optimization algorithm.

[0046] (5) Convergence of optimization results and selection of optimal parameter combinations. Repeat step (4) until the overall optimal fitness function becomes stable and meets the convergence criteria. For single-objective optimization problems, the optimal parameter combination can be directly obtained; for multi-objective optimization problems, it is necessary to further weigh the priorities between the objectives and then determine the optimal parameter combination. Using the parameter mapping method, the optimal parameter combination is mapped to the structural information model to obtain the optimized structural information model, thereby realizing the rapid design of beam-slab steel structures.

[0047] Compared with the prior art, the present invention has the following advantages:

[0048] The present invention realizes efficient modeling, comprehensive information characterization and dynamic updating of beam-slab steel structures through a modeling method based on a graph data structure and a parametric characterization mechanism. Compared with traditional two-dimensional design and three-dimensional modeling, the present invention can more intuitively express complex three-dimensional spatial relationships, avoid information transmission errors, and support rapid modification and updating, which significantly improves design efficiency and quality. Its powerful parametric characterization function makes structural design more flexible, and can quickly generate and modify large quantities of models to meet the refined design requirements of modern complex structures. In addition, the present invention ensures the accuracy and integrity of the structural design by automatically searching and verifying directed closed loops. It is widely applicable to fields such as marine engineering, construction engineering and transportation engineering, and can effectively respond to the design and construction needs of large and complex beam-slab steel structures, and has important practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a schematic diagram of the steel structure of the transition section beam plate of the offshore wind turbine jacket;

[0050] Figure 2 Schematic diagram of the steel structure of the transition beam plate of the offshore wind turbine jacket represented by a graph data structure. Specific implementation methods

[0051] The present invention will be further described below with reference to the accompanying drawings and examples. This example is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following examples.

[0052] The information modeling and parameterized characterization method of the beam-slab steel structure of the present invention includes:

[0053] 1. The specific steps of information modeling of beam-slab steel structure based on graph data structure are as follows:

[0054] (1) Abstract components into graph data structures. Abstract the components (beams, plates, columns, connectors, etc.) in beam-slab steel structures into spatial nodes (vertices), line segments (edges), and faces (directed closed loops) in graph data structures. Vertices correspond to nodes of the structure, recording their spatial coordinates and connection attribute information sets; edges correspond to component line segments, recording line types, control parameters, and component attribute information sets; directed closed loops correspond to structural surfaces, recording surface types, thicknesses, and material attribute sets. The specific steps are as follows:

[0055] (2) Vertex modeling: All key nodes in the design graph (assuming a total of n vx nodes), modeled as vertices using the following expressions:

[0056]

[0057] Where, v i is any vertex in the space, and V is the set of all vertices in the structure. i ,y i ,z i is the spatial coordinate of the vertex. For vertex v i As a component attribute information set when connecting nodes.

[0058] (3) Edge modeling: Model all line segments in the design diagram. The edge expression is:

[0059]

[0060] Where, e ij To connect the vertex v i and v j The edges of , E is the set of all edges in the structure, is a set of structural surface labels, indicating e ij One or more tags are located at ET1 to On the surface; ES ij Indicates the line type of the edge; P ij ={p1,p2,...} are parameters that control the shape of the curve and are mainly used to define ellipses and Bezier curves. For edge e ij A set of component property information when used as a structural component (beam, column).

[0061] (4) Modeling and searching for directed closed loops. The expression of a directed closed loop is:

[0062]

[0063] Where c k Represents the kth directed closed loop in the graph, consisting of a series of edges Composition, subscript i i ...i n Refers to the vertices of the edge, where the starting and ending vertices are the same. In addition to the defined edges, the loop has two other structural construction properties: CE k is the surface type, where the integer CE k =1,2 represent plane surface and curved surface respectively; t k is the thickness property of the kth surface / directed loop. is a directed closed loop c k A set of component attribute information when representing a structural surface.

[0064] The Tarjan strongly connected component algorithm is used to automatically search for directed closed loops in the graph, and the legitimacy of the loops is verified by the intersection of thickness sets. Only loops that satisfy the following expression are retained as valid structural surfaces:

[0065]

[0066] The subscript of ET in the formula represents the vertex number of the corresponding edge. It is necessary to ensure that the first vertex number of the first edge is consistent with the vertex number of the last edge, that is, i1. The loop that does not meet the above check will not be included in the depth-first search stack, and the loop that passes the check will be assigned a component attribute information set

[0067] 2. Definition and parameterization of component attribute information set. The specific steps are as follows:

[0068] (1) Definition of Vertex Attribute Information Set. Vertex Attribute Information Set Include connection type identification variable like Indicates no connection. Indicates welding, Indicates bolt connection and corresponding welding parameters Bolt parameters and material properties

[0069] (2) Definition of edge attribute information set. Include cross-section type identification variable (such as standard section or custom section), section parameters (such as I-section model or polyline endpoint coordinates of custom section), thickness t k and material properties

[0070] (3) Definition of directed closed-loop attribute information set. Directed closed-loop attribute information set Including surface thickness t k , internal boundary set c in , and material property parameters If there is a hollow, the thickness of the internal boundary directed closed loop is set to 0, and the material property set is an empty set.

[0071] 3. Design parameter mapping representation. The specific steps are as follows:

[0072] (1) Geometric parameter mapping. Determine the design parameter set based on the design drawings. (such as length, angle, curvature, etc.), manually establish the mapping function Associate geometric parameters with vertex coordinates to achieve parameter-driven dynamic adjustment of geometric shapes.

[0073] (2) Linear and surface parameter mapping. Determine the line shape of the edge according to the design drawing ij and surface type ST k and through the control parameter P ij (such as the endpoints of the semi-major axis of an ellipse and the control points of a Bezier curve) to accurately describe the shape of the curve.

[0074] (3) Attribute parameter mapping.

[0075] Based on the standard section library and material library, the section identification and edge attribute information in the design drawing are combined Surface attribute set Association; connect node information with vertex attribute information set Association, ensuring the consistency of parameterized representation.

[0076] 4. Model verification and dynamic update. The specific steps are as follows:

[0077] (1) Model integrity verification. The Tarjan algorithm is used to automatically verify the closure and thickness consistency of the directed closed loop, and loops that do not meet the conditions are eliminated to ensure the physical rationality of the structural surface.

[0078] (2) Dynamic update of parameters. When the component properties (such as section type, material) change, the vertex coordinates, edge shape parameters and closed-loop thickness are automatically updated through the mapping function to achieve real-time dynamic adjustment of the model.

[0079] 5. Structural optimization, the specific steps are as follows:

[0080] (1) Establish a single-objective or multi-objective optimization problem, including the optimization objective (minimization of mass or minimization of structural response) and constraints (limit condition constraints, fatigue condition constraints, buckling condition constraints, etc.).

[0081] (2) Determine the range of design parameter values. Determine the reasonable range of each design parameter based on the design drawings. Use the parameter mapping method to build an information model. Read the geometry and component information based on the 3D modeling software script data interface to perform geometric modeling and divide the finite element mesh.

[0082] (3) Solve the fitness function of the optimization problem. Using the finite element mesh generated in step (2), apply boundary conditions and design loads, perform finite element simulation, read key structural response indicators, and automatically calculate the overall mass based on the structural information model. Construct the fitness function based on the optimization problem description in step (1).

[0083] (4) Update the design parameters. Use a discrete value swarm optimization algorithm (such as genetic algorithm, ant colony algorithm, particle swarm algorithm, etc.) to calculate steps (1)-(3) for each individual, integrate the fitness function results, and update the design parameters according to the optimization algorithm strategy.

[0084] (5) Convergence of optimization results and selection of the optimal parameter combination. Repeat step (4) until the fitness function is stable and meets the convergence criteria. Single-objective optimization can directly determine the optimal parameter combination, while multi-objective optimization requires weighing the priority of the objectives to determine the optimal combination. Through parameter mapping, the optimal parameter combination is mapped to the structural information model to complete the rapid optimization design of the beam-slab steel structure. Example

[0085] Given an offshore wind turbine jacket transition section structure, such as Figure 1 shown. Figure 1 In (b), 1 is the main tube, which is the main load-bearing structure. 2 is the inclined support box beam, which connects the top of the jacket and the main tube. 3 is the bottom plate, which needs to be reinforced with a reinforcement beam to prevent buckling deformation. Figure 1 -(c). 4 is the top cover plate, which has a more complex shape. Figure 1 -(c) In the bottom plate, bottom plate reinforcement plates are arranged at the four corners of the bottom plate (such as Figure 1 (as shown in 6).

[0086] Example: Figure 1A total of nine nodes are marked, labeled P1-P9. In addition, seven line segments are labeled E23-E94. Line segments E45-E94 form a directed loop, representing a structural surface S-1. The unreinforced surface of the base plate is labeled S-2.

[0087] 1. First, construct a graph data structure and perform information modeling on the transition segment structure.

[0088] (1) Vertex modeling, Figure 1 Take point P-1 in the example, its coordinates are (0, 2500, 4000), the length unit is millimeter, and the number is 1, then its vertex information is:

[0089]

[0090] (2) Edge modeling Figure 1 Take E-56 in the example, its connecting nodes P-5 and P-6, represented by a line segment of a quarter arc, then its edge information is:

[0091]

[0092] Among them, the structural surface label ET 56 = {S1, S2}, indicating that edge E-56 is on surfaces S-1 and S-2. Linear variable ES = 2, indicating that the line segment belongs to the generalized ellipse segment. Therefore, the curve shape control parameter set P 56 ={-4400, 4400, 0, 100, -4500, 4400, 0}, where the first three digits are the coordinates of the center of the circle, the fourth digit is the radius of the major axis (for an arc, that is, the radius), and the last three digits are the coordinates of the major axis point.

[0093] (3) Directed closed-loop modeling Figure 1 Take the total S-1 as an example, which is composed of edges E45-E94, and its expression is:

[0094]

[0095] The surface type variable CE1=1, indicating that S-1 is a plane.

[0096] The Tarjan strongly connected component algorithm is used to automatically search for directed closed loops in the graph, and the loop legitimacy is verified by thickness set intersection. For the structural surface S-1, the following verification is performed:

[0097] ET 45 ∩ET 56 ∩ET 67 ∩ET 78 ∩ET 89 ∩ET 94 ={S1}

[0098] Indicates that all edges on a loop share a structural face label.

[0099] 2. Define the definition of component attribute information set.

[0100] (1) Definition of vertex attribute information set. Taking node P-2 (vertex v2) as an example, it is a welding node, so its information set is Defined as:

[0101]

[0102] Among them, the second to fourth parameters represent the weld length (mm), weld effective thickness (mm), weld angle (°) and weld groove form respectively. The first parameter indicates the base material type, and the second parameter is the welding material model.

[0103] (2) Definition of edge attribute information set. Take line segment E-23 as an example, it is 10# channel steel, so its information set for:

[0104]

[0105] Among them, the material information set is

[0106] (3) Definition of directed closed loop attribute information set. Taking the structural surface S-1 as an example, its attribute information set The expression is:

[0107]

[0108] Among them, 60 is the thickness of the structural surface, in millimeters. The inner boundary closed loop set is an empty set, indicating that there is no internal structural surface or hollowing. The material information set is

[0109] 3. Design parameter mapping representation relationship definition.

[0110] (1) Forming structural geometric design parameters Directly with the vertex v i The spatial coordinate x i ,y i ,z i Mapping relationship. Taking node P-5 as an example, its coordinates are affected by the length of the bottom plate l b =9000mm and the outer arc radius r of the bottom plate b =100mm control, so the design parameters can be expressed as So the expression of the mapping relationship is:

[0111]

[0112] (2) Standard library establishment. The standard libraries involved in this model are steel standard sections and material standard libraries. The material standard sections mainly include standard sizes of I-type and C-type steel, with reference to the national standard GB / T 706-2016. The material standard library mainly includes steel types and welding material types, with reference to European standard EN 10025, national standard GB / T 700-2006, American Welding Society standard AWS A5.18, and national welding standard GB / T 985.1-2018, etc.

[0113] Through the above three steps, a complete graph data structure representing the jacket transition section beam plate structure can be obtained, such as Figure 2 shown.

[0114] 4. Model verification and dynamic update.

[0115] (1) After completing the overall information model construction and parameterized representation, it is necessary to use the Tarjan strongly connected component algorithm again to automatically search for directed closed loops in the graph data structure to ensure that there are no duplicate definitions or invalid definitions.

[0116] (2) If the model needs to redefine or modify a parameter, only the parameters in the parameter set need to be modified. For example, if the length of the bottom plate l needs to be modified b =9000→l b =8000, then all relevant points (the input parameter subset in the mapping function includes the variable of the bottom plate length, i.e. ) will automatically update the information, such as point P-5 information will become:

[0117]

[0118] 5. Structural optimization.

[0119] (1) Construct an optimization problem with the single objective of minimizing the structural mass m and constraining the maximum stress σ of the structure. max Less than 300MPa, fatigue strength of welds at key locations f is less than 1, so the optimization problem can be written as:

[0120] Min. m(x)

[0121] ST σ max (x)<300

[0122] max(D f (x))<1

[0123] The fitness function expression of the optimization problem is constructed based on the penalty function method:

[0124] F(x)=m(x)+λ1max(σmax (x)-300,0)+λ2max(max(D f (x))-1,0)

[0125] Where λ1 and λ2 are penalty coefficients, and λ1=λ2=10 6 .

[0126] (2) Determine the range of design parameters, such as the length of the base plate l b ∈[8000,9000], with an interval of 100 mm. Based on the parameter mapping method and the third-party modeling software interface, the geometry and component information are read, the geometric modeling is performed, and the finite element mesh is divided using the triangulation algorithm.

[0127] (3) Solve the fitness function of the optimization problem. Using the finite element mesh generated in step (2), apply the bottom support, the top wind load, and the weight of the upper structure. Automatically calculate the overall mass m based on the structural information model, perform static finite element simulation, and read the maximum stress σ max Use equivalent fatigue load method to perform fatigue verification and read the maximum fatigue damage D f Integrate the above variables to solve the fitness function in step (1).

[0128] (4) Update the design parameters. Use the genetic algorithm to calculate steps (1)-(3) on the individuals, integrate the fitness function results, perform crossover and mutation on the individuals, and update the design parameters.

[0129] (5) Convergence of optimization results and selection of optimal parameter combination. Repeat step (4) until the minimum fitness function of the group remains unchanged in 50 iterations, and the final optimal parameter combination of the group is the optimal parameter combination. There is randomness in the optimization process, so there are differences in the optimization results each time. This embodiment makes three optimization attempts in total to avoid possible premature convergence problems. The overall mass of the results obtained from the three optimization attempts is m1(x) = 452.3 tons, m2(x) = 447.6 tons and m1(x) = 434.1 tons, respectively, which is a significant decrease compared to the initial design m0(x) = 487.2 tons. The optimal parameter combination is mapped to the structural information model through parameter mapping to complete the rapid optimization design of the jacket transition section structure.

Claims

1. A beam-slab steel structure optimization design method based on information modeling and parametric characterization, characterized in that: The following steps are involved: (1) Based on the graph data structure, the information model of the beam-slab steel structure is obtained; (2) mapping the design parameters to the structural information model obtained in step (1) to obtain a structural information basic model containing the initial design parameters; (3) Based on the structural information basic model containing the initial design parameters obtained in step (1), the design parameters are adjusted and optimized through an automatic optimization algorithm to finally obtain an optimized structural information model. The beam-slab steel structure is optimized according to the optimized structural information model.

2. The beam-slab steel structure optimization design method based on information modeling and parametric characterization according to claim 1 is characterized in that: In step (1), the information modeling of the beam-slab steel structure based on the graph data structure is used to obtain the information model of the structure, which specifically includes: Each component in the beam-slab steel structure is abstracted into structural nodes, structural segments and structural surfaces. The vertices in the graph data structure are used to perform information modeling on the structural nodes, and the edges in the graph data structure are used to perform information modeling on the structural segments. The Tarjan strongly connected component algorithm is used to search for directed closed loops in the graph data structure, and the directed closed loops are used to perform information modeling on the structural surfaces. Finally, different component attribute information sets are defined in the graph data structure to obtain the information model of the beam-slab steel structure.

3. The beam-slab steel structure optimization design method based on information modeling and parametric characterization according to claim 2 is characterized in that: Use the vertices in the graph data structure to perform information modeling on the structural nodes, including: (1.1) Vertices are used to perform information modeling on structural nodes, and the expression is: Where, v i is any vertex of the structure, V is the set of all vertices in the structure, a total of n v vertices, x i ,y i ,z i is the spatial coordinate of this vertex, For vertex v i As a component attribute information set when connecting nodes.

4. The beam-slab steel structure optimization design method based on information modeling and parametric characterization according to claim 2 is characterized in that: Use the edges in the graph data structure to perform information modeling on the structural segments, including: (1.2) Use the edge to perform information modeling on the structural line segment, and the expression is: Where, e ij To connect the vertex v i and v j The edges of , E is the set of all edges in the structure, is a set of structural surface labels, indicating e ij One or more tags are located at ET1 to On the surface; ES ij Indicates the line type of the edge; P ij ={p1,p2,...} are the parameters that control the shape of the curve, For edge e ij A set of component property information when it serves as a structural component.

5. The beam-slab steel structure optimization design method based on information modeling and parametric characterization according to claim 4 is characterized in that: The Tarjan strongly connected component algorithm is used to search for directed closed loops in the graph data structure and to perform information modeling on the structural surface using directed closed loops, specifically including: Based on the side information model e obtained in step (1.2) ij , the Tarjan strongly connected component algorithm is used to automatically search for directed closed loops in the graph data structure. A directed closed loop in the graph data structure represents only one structural surface, and there is no overlap between any two directed closed loops. A secondary check is performed on each searched directed closed loop to ensure that the intersection of the structural surface label sets of all edges on the loop of the directed closed loop is not an empty set, that is: The above formula shows that all edges of the directed closed loop share a thickness label; A directed closed loop is used to perform information modeling on the structural surface, and its expression is: Where c k Represents the kth directed closed loop in the graph, consisting of a series of edges Composition, subscript i i ...i n Refers to the vertex label of the edge, C is the set of all directed closed loops in the structure, CE k is the surface type; is a directed closed loop c k A set of component attribute information when representing a structural surface.

6. The beam-slab steel structure optimization design method based on information modeling and parametric characterization according to claim 2 is characterized in that: Define different component attribute information sets in the graph data structure, including: For vertex v i Component attribute information set Its expression is: Where, Identify the variable for the connection type, represents an unconnected free vertex, Only identification variables are included in Represent welding, bolt connection and bolt-weld connection respectively. It is the welding related information parameter, It is the bolt connection related information parameter, is the material property set for the connection; For edge e ij Component attribute information set Its expression is: Where, Identify the variable for the beam's cross-section type, Represents edge e ij Not used as a structural member; Indicates the standard cross-section form, Represents different standard sections, and the relevant parameter sec is the model of the standard section. Indicates a non-standard section defined by a polyline, p x,k and p y,k Indicates the local coordinates of the endpoints of the section polyline, t k is the thickness at the end point of the kth section, and the section definition has n edge endpoints, is the material property set for the edge; For a directed closed loop c k Component attribute information set Its expression is: Where, t k represents the thickness set of the panel boundary nodes, c in is the internal closed loop set, representing the inner boundary of the panel, c in When it is an empty set, it means that the panel has no internal boundary, c in Containing element 0 indicates forming a partial hollowing. is the material property set of a directed closed loop. When the closed loop represents an internal hollow, 7. The beam-slab steel structure optimization design method based on information modeling and parametric characterization according to claim 1 is characterized in that: In step (2), the design parameters are mapped to the structural information model obtained in step (1) to obtain a structural information basic model containing the initial design parameters, specifically including: (1) Determine the design parameter set based on the design drawings Establish a mapping function to transform the design parameter set Associated with the spatial coordinates of the vertices, through calculation and optimization, adjust the design parameters to form the structural geometric design parameters Directly with the vertex v i The spatial coordinate x i ,y i ,z i Mapping relationship, mapping relationship f i The expression is written as: (2) Determine linear ES based on design drawings ij , and according to ES ij Determine the control parameter set P of the corresponding edge ij , for circular arcs, P ij Contains the coordinates of the circle center; for elliptical arcs, P ij Contains the coordinates of the center point and the semi-major axis endpoints; for Bezier curves, P ij Contains the number of control points, control point coordinates and curve control parameters, and defines the edge according to the shape of the structural component ij The linear parameter ES in ij and P ij ; (3) Determine the surface type CE according to the design drawing k And the thickness set t of the panel boundary nodes under the corresponding parameter value k , determine whether there is a loop inside the face to construct the internal closed loop set c in ; (4) Establish a standard library of nodes and beam sections, assign standard numbers, build a standard library of steel structure materials, and set the material properties of structural nodes, segments and surfaces. and Select and define from the standard material library, and determine the attribute information set based on the cross-section drawing of the design drawing and (5) According to steps (1), (2), (3) and (4), a structural information basic model including initial design parameters is formed.

Images