A parameterized modeling and optimization design method for a portal crane box girder

Abstract

The application discloses a portal bridge crane box-shaped main beam parameterized modeling and optimization design method, and belongs to the technical field of mechanical design, which comprises the following steps: realizing parameterized geometric modeling of the main beam of the crane in a three-dimensional modeling software; automatically importing a geometric model output by the three-dimensional design software into a finite element pre-processing software, and performing modeling pre-processing on the geometric model to establish a parameterized finite element simulation model of the main beam of the crane; performing pre-processing and solution calculation on the finite element simulation analysis of the main beam structure of the parameterized finite element simulation model; establishing an approximate proxy model of the parameterized finite element simulation model based on DOE analysis of the finite element; and using an optimization algorithm to perform optimization under safety and manufacturability constraints for the approximate proxy model on the basis of the parameterized finite element simulation model. Through the application, simple parameterized finite element modeling of the design of the main beam of the crane box can be realized, and the most economical lightweight crane structure size can be quickly found.

Description

Technical Field

[0001] This invention belongs to the field of mechanical design technology, and more specifically, relates to an optimization design method for the metal structure of mechanical equipment, particularly the box girder structure of a gantry crane. Background Technology

[0002] Cranes are equipment used for lifting, handling, loading, and unloading materials. The design and calculation of the metal structure of gantry cranes inevitably involves statically indeterminate spatial structures. Furthermore, the calculations involve numerous operating conditions, making manual calculations insufficient to handle the complex analyses and heavy workloads. Traditional design calculation methods necessitate various simplifications and assumptions. While these simplifications make the calculation process feasible and concise, they also result in coarse calculations that deviate significantly from reality, requiring compensation through increased safety factors. This leads to larger structural dimensions, material waste, increased weight, higher energy consumption, and increased manufacturing, transportation, plant construction, and operating costs. Currently, the structural dimensions of gantry cranes designed and used in China are generally larger than those of comparable tonnage products from abroad.

[0003] Traditional main beam structural optimization design first relies on experience and judgment to select and determine the structural scheme, initially choose the cross-sectional dimensions of the components, and then perform verification calculations for strength, stiffness, and stability. Modifications to the scheme or comparisons of a limited number of schemes are also verification-based. Due to the enormous computational workload, only a small number of schemes can be compared, and the quality of the structural design depends excessively on the designer's skill and experience, making it difficult to arrive at a satisfactory solution.

[0004] The development of computer CAE and optimization technology has provided new methods for the optimized design of box girder main beams of gantry cranes.

[0005] 1) Some optimization techniques (such as fruit fly algorithm, moth-flame algorithm, genetic algorithm, response surface optimization, bee colony optimization, etc.) are employed, but the nominal stress method of traditional engineering mechanics is still used for calculations in order to overcome the drawbacks of manually defined design parameters and achieve parameter optimization. The formula method of engineering mechanics involves some simplifications and assumptions, generally only considering the upper and lower cover plates and web plates, without fully considering the stress and stability safety effects of geometric discontinuities and complex structures such as diaphragms and stiffeners. The calculation results cannot reflect the actual structural condition, leading to poor optimization results.

[0006] 2) Some designs also employ the aforementioned optimization techniques combined with the finite element method for multi-objective optimization of crane main beams. However, the parametric and automated modeling optimization process of the finite element model is complex, and effective quality control is lacking in the preprocessing of the finite element model. This results in the connections between components not reflecting the actual connection situation and poor finite element mesh quality control. Most designs rely on the internal parameter control of the finite element software for automatic mesh generation, leading to low mesh quality, localized stress concentration, and low calculation accuracy, thus affecting the optimization effect of the crane main beam.

[0007] In summary, existing classic designs for gantry crane box girder main beams are based on traditional engineering mechanics, making numerous assumptions and simplifications for ease of calculation, resulting in significant discrepancies between calculated results and actual conditions. Furthermore, due to the numerous operating conditions and substantial computational load, structural optimization relies excessively on the designer's skill and experience, typically involving only a limited number of comparisons, making it difficult to arrive at a satisfactory solution. Existing design methods for gantry crane box girder main beams using modern optimization techniques combined with the finite element method generally lack effective quality control during CAE preprocessing, resulting in poor mesh quality, localized stress concentrations, and low computational accuracy, thus affecting the optimization effect of the crane main girder. Summary of the Invention

[0008] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention proposes a parametric modeling and optimization design method for the box girder of a gantry crane. This method enables simple parametric finite element modeling of the crane's box girder, effectively controls preprocessing quality, improves analysis accuracy, and utilizes optimization algorithms to optimize the design parameters of the crane's box girder based on the parametric finite element model and the approximate model constructed through DOE analysis, quickly finding the most economical lightweight crane structural dimensions.

[0009] To achieve the above objectives, the present invention provides a parametric modeling and optimization design method for the box girder of a gantry crane, comprising:

[0010] Implement parametric geometric modeling of the crane's main beam in 3D modeling software and output the geometric model;

[0011] The geometric model output by the 3D design software is automatically imported into the finite element preprocessing software, and the geometric model is preprocessed to establish a parametric finite element simulation model of the crane main beam.

[0012] Preprocessing is performed on the parametric finite element simulation model of the main beam of the crane for finite element simulation analysis, and the preprocessed parametric finite element simulation model is solved and calculated.

[0013] An approximate proxy model for the parametric finite element simulation model of the crane main beam is established based on DOE analysis of finite element method.

[0014] Based on the parametric finite element simulation model of the crane main beam, optimization algorithms are used to find the best model under safety and manufacturability constraints for the approximate surrogate model constructed by DOE analysis.

[0015] In some optional implementations, the parametric geometric modeling of the crane main beam in 3D modeling software and the output of the geometric model include:

[0016] The geometric models of the upper and lower flange plates, web plates, diaphragm plates, end plates, and stiffening ribs of the box girder of the gantry crane were established using 3D modeling software. A solid 3D model of the track was also established. The geometric model plates were modeled using the mid-surface, and geometric features with little impact on the calculation were ignored. The width of the upper and lower flange plates, the height of the web plates, the spacing between the web plates, the number of intermediate diaphragm plates, and the spacing of the stiffening ribs of the main girder were parameterized to form a parameter table. The geometric model was output in a file format supported by the finite element preprocessing software.

[0017] In some optional implementations, the automatic import of the geometric model output by the 3D design software into the finite element preprocessing software, and the preprocessing of the geometric model, includes:

[0018] The area of ​​the mid-surface geometry is calculated in the 3D modeling software and compared with the area of ​​the mid-surface geometry calculated in the finite element software to determine the correctness of the geometric model output from the 3D modeling software to the input of the finite element preprocessing software. This is used to perform preprocessing on the geometric model until the geometric model in the finite element preprocessing software is consistent with the geometric model in the 3D modeling software, thus obtaining the parametric finite element simulation model of the crane main beam.

[0019] In some optional implementations, the preprocessing for the finite element simulation analysis of the parametric finite element simulation model of the crane main beam includes:

[0020] The components of the parametric finite element simulation model are connected according to welding and bolting relationships to simulate actual weld connections. Utilizing the relatively fixed ID numbers and component names of the imported geometric model elements, geometric subdivision is performed on the intersection surfaces of the web, upper and lower flanges, and diaphragms based on the ID numbers of geometric points, lines, and surfaces. Mesh generation and smoothing are then performed. The main beam components use shell elements, while the trolley track uses solid elements. Material and section properties are assigned to elastoplastic bodies, and loads are set, including load steps, load combinations, and boundary constraints. The parametric finite element simulation model undergoes weight verification. The total weight of the parametric finite element simulation model is calculated and compared with the total design weight of the main beam to check the correctness of the mesh, material properties, and section property settings. If the total weight of the parametric finite element simulation model is not equal to the total design weight of the main beam, the mesh, material data, and section property settings are checked and corrected.

[0021] In some optional implementations, the solution calculation of the preprocessed parametric finite element simulation model includes:

[0022] The pre-processing process of the main beam structure is further developed to form a standardized processing script program. The plate thickness of each component of the main beam is set in the script program. The script program automatically imports the geometric model file, performs connection, mesh, load, constraint, and load step processing, and outputs the finite element solution file.

[0023] Import the solution file into the finite element software solver for calculation, and extract the strength, stiffness and stability results from the result file to output the weight of the parameterized finite element simulation model.

[0024] In some optional implementations, the finite element-based DOE analysis, which establishes an approximate proxy model for the parametric finite element simulation model of the crane main beam, includes:

[0025] For key parameters such as dimensions and thickness of components within the design space, sampling is performed according to the DOE (Design of Experiments) experimental design method. The geometric model is automatically modified based on the changing parameters. A standardized pre-processing script is used to quickly generate a parametric finite element simulation model and output a solution file. The finite element solver is then called to calculate the output results. Strength, stiffness, stability, and weight information are extracted to form sample point data. Based on the design parameter variables and the response data from the calculation results, an approximate surrogate model is established using a model fitting algorithm. The goodness-of-fit R-squared of the surrogate model is then evaluated. 2 Error analysis and evaluation were conducted to obtain an approximate surrogate model that meets the requirements.

[0026] In some alternative implementations, the goodness-of-fit R-value of the surrogate model is... 2 Error analysis and evaluation yielded an approximate surrogate model that meets the requirements, including:

[0027] The goodness-of-fit R-value of the surrogate model is calculated. 2 Error analysis and evaluation, if R 2 If the result is greater than 0.9, proceed to the next step; otherwise, adjust the fitting parameters of the surrogate model, change the fitting algorithm, or add sample data to refit until the goodness-of-fit requirement is met, thus obtaining a satisfactory approximate surrogate model.

[0028] In some optional implementations, the optimization algorithm, based on the parametric finite element simulation model of the crane main beam, performs optimization under safety and manufacturability constraints on the approximate surrogate model constructed through DOE analysis, including:

[0029] Using parametric dimensions as design variables, the design space is limited by actual production and manufacturing conditions, with strength, stiffness, stability and manufacturability indicators as constraints, and the weight of the main beam as the optimization objective. An intelligent optimization algorithm is adopted to iteratively optimize based on a surrogate model.

[0030] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0031] (1) This invention provides a method for parametric modeling and optimization design of the box girder of a gantry crane. Parametric geometric modeling of the crane girder is achieved in 3D design software. The geometric files output by the 3D design software are automatically imported into the finite element preprocessing software to establish a parametric finite element analysis model of the crane girder, effectively controlling the preprocessing quality and improving analysis accuracy. Based on the parametric finite element model, optimization algorithms are used to optimize the approximate model constructed through DOE analysis under safety and manufacturability constraints such as strength, stiffness, and stability. This optimizes the dimensions and material thickness of the crane girder, reduces material usage, achieves lightweighting, and improves product economy.

[0032] (2) This method enables accurate and rapid modeling of the parametric finite element simulation model of the box girder of a gantry crane, preserving the components of the girder to the greatest extent possible, reflecting the actual structural form, and considering its impact on design performance. Quality control is implemented in the preprocessing of component connections, meshes, and loads during modeling, improving the overall analytical accuracy of the finite element simulation model. It reduces problems such as improper connections and stress concentration caused by coarse model processing, allowing for a more realistic reflection of the actual situation.

[0033] (3) Based on the parameterized precise finite element model, optimization algorithms are used to perform large-scale optimization within the design space under constraints such as design performance requirements and manufacturability. This overcomes the problems of simplification due to complex calculations, coarse results caused by assumptions, oversized crane structural dimensions, and material waste caused by traditional methods, thereby optimizing the material usage of the crane main beam and improving economic efficiency. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating a parametric modeling and optimization design method for a box girder of a gantry crane provided in an embodiment of the present invention.

[0035] Figure 2 This is one of the embodiments provided by the present invention. Figure 1 A step-by-step diagram of step 3 in the middle section;

[0036] Figure 3 This is a schematic diagram of key structural parameters of a main beam provided in an embodiment of the present invention;

[0037] Figure 4 This is one of the embodiments provided by the present invention. Figure 1 A flowchart illustrating step 7. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0039] In the embodiments of the present invention, "first," "second," etc. (if present) are used to distinguish different objects, rather than to describe a specific order or sequence.

[0040] like Figure 1 The diagram shown is a flowchart illustrating a parametric modeling and optimization design method for a box girder of a gantry crane according to an embodiment of the present invention, which includes the following steps in sequence:

[0041] Step (1): Perform three-dimensional geometric parametric modeling of the crane box girder;

[0042] Three-dimensional modeling software (such as CATIA, SolidWorks, UG, etc.) was used to create geometric models of the main beam's upper and lower flange plates, web plates, diaphragms, end plates, stiffeners, and other components based on their mid-surfaces. Solid 3D models of the tracks were also created. To facilitate subsequent finite element model preprocessing, the geometric model plates were modeled only based on their mid-surfaces, ignoring geometric features with minimal impact on calculations: for example, the fillets at the edges of plates and the fillets on bent surfaces were treated as right angles, and holes with minimal force impact were removed. The key dimensions and assembly position dimensions of the main beam components were parameterized, such as the width of the upper and lower flange plates, web plate height, web plate spacing, number of mid-section diaphragms, and stiffener spacing, forming a parameter table.

[0043] Write macro command scripts to modify and update the parameters of each component of the main beam, such as key dimensions and assembly position dimensions, and finally output the geometric model in a geometric file format supported by the finite element preprocessing software.

[0044] Step (2): Import the geometric file into the finite element preprocessing software;

[0045] In the 3D modeling software, the area A1 of the mid-surface geometry is calculated, and in the finite element software, the area A2 of the mid-surface geometry is calculated. It is determined whether A1 is equal to A2 to confirm the correctness of the geometric model output from the 3D modeling software to the input of the finite element preprocessing software. Errors such as missing or damaged surfaces are checked. If A1≠A2, the geometric model is corrected in step (1), and the 3D geometry and file output parameters are adjusted until the geometric model in the finite element preprocessing software is consistent with the geometric model in the 3D modeling software.

[0046] Step (3): Preprocessing for finite element simulation analysis of the main beam structure;

[0047] like Figure 2 As shown, the model components are connected according to welding, bolting, and other relationships to simulate actual weld connections: intersecting surfaces are extended to achieve mesh common nodes on the intersecting lines; the welding of the upper and lower contact edges of parallel surface components is handled by overlapping surface thicknesses, merging geometric surfaces, and taking advantage of the relatively fixed characteristics of the imported geometric model's geometric element IDs and component names, geometric segmentation is performed on the intersecting surfaces of the web, upper and lower flanges, and partitions based on the IDs of the geometric model's points, lines, and surfaces. The mesh is then divided and smoothed, and the mesh quality is checked according to the element quality requirements in the "General Rules for Finite Element Mechanical Analysis of Mechanical Product Structures". Mesh that does not meet the requirements is split into surfaces and geometrically processed until the mesh quality meets the requirements.

[0048] The main beam components are treated with plate and shell elements, while the trolley track is treated with solid elements. Material and section properties are assigned to the elasto-plastic bodies, and thickness information is assigned to the plate and shell elements in the section properties. Loads are set according to the national standard "Crane Design Code" and design requirements, and load steps, load combinations, and boundary constraints are established. The wheel pressure of the trolley is coupled to the corresponding position on the track in the form of concentrated forces.

[0049] The model is weight-checked by calculating the total weight M1 of the finite element model and comparing it with the total weight M2 of the main beam design to check the correctness of the model mesh, material properties, and section property settings. If M1 ≠ M2, the mesh, material data, and section property settings are checked and corrected.

[0050] Step (4): The pre-processing process of the main beam structure is further developed to form a standardized processing script program;

[0051] The preprocessing process in step (3) uses script commands to form a standardized processing flow program. The thickness of each component of the main beam is set in the script to facilitate subsequent parameterization. The script program automatically imports the geometric model file, performs connection, mesh, load, constraint, and load step processing, and outputs the finite element solution file.

[0052] Step (5): Calculate using the finite element solution file and output the calculation results for strength, stiffness, stability, etc.

[0053] Import the solution file into the finite element software solver for calculation, and extract results such as strength, stiffness, and stability from the result file, while also outputting the weight information of the model.

[0054] Step (6): Based on finite element DOE analysis, establish an approximate proxy model for the simulation model;

[0055] like Figure 3 As shown, for the parameters of the above components, including key parameters such as size and thickness, sampling is performed within the design space area according to the DOE experimental design method. The macro script in step (1) automatically modifies the three-dimensional geometric model according to the changing parameters. The preprocessing standardized script program in step (4) quickly generates the finite element simulation model and outputs the solution file. The finite element solver is called to calculate the output results. The results of strength, stiffness, stability, and weight information are extracted to form sample point data. Based on the design parameter variables and the response data of the calculation results, an approximate surrogate model is established using a suitable model fitting algorithm (including but not limited to RBF neural network, Kriging, and RSM response surface). The goodness-of-fit R of the surrogate model is then evaluated. 2 Error analysis and evaluation, such as R 2If the result is greater than 0.9, proceed to the next step; otherwise, adjust the fitting parameters of the surrogate model, change the fitting algorithm, or add sample data to refit until the goodness-of-fit requirement is met, at which point the approximate surrogate model can be used.

[0056] Step (7): Optimize the main beam of the crane based on the approximate surrogate model that meets the requirements;

[0057] like Figure 4 As shown, the parameterized dimensions in the aforementioned steps are used as design variables. The design space is limited based on the actual production and manufacturing conditions. Indicators such as strength, stiffness, stability, and manufacturability are used as constraints. The weight of the main beam is used as the optimization objective. Modern intelligent optimization algorithms (including but not limited to genetic algorithms, sequential quadratic programming methods, and their combinations) are used to iteratively optimize based on a surrogate model.

[0058] Example: The following example provides a specific method for parametric modeling and optimization design of the box girder of a 20t 22.5m general-purpose bridge crane.

[0059] Step (1): Perform 3D geometric modeling of the crane box girder. Use CATIA 3D modeling software to create geometric models of the upper and lower flanges, webs, diaphragms, end plates, stiffening ribs, and other components of the main girder using mid-surface geometry, and create solid 3D models of the track. Create part drawings of the above plates containing only mid-surface geometry in the "Generative Shape Design" workbench of CATIA.

[0060] The rounded corners of the plate edges and the rounded corners of the stiffening rib bending surfaces are treated as right angles, and holes with minimal impact on force are removed. The width of the upper and lower flange plates of the main beam, the height of the web plate, the web plate spacing, the number of intermediate partition plates, and the spacing of the stiffening ribs are parameterized to form a parameter table.

[0061] The parameterization method for the number of partition plates is as follows: 1) In the "Generative Shape Design" workbench interface of CATIA, create point dot1 on the edge line of the upper cover plate along its length direction. The distance of dot1 relative to the edge of the upper cover plate is the distance between the first full-height partition plate on the edge and the end. 2) Perform a "rectangular array" on dot1 along the edge line of the upper cover plate, and parameterize the number of instances to "number of partition plates". The array spacing = total spacing of partition plates / (number of instances - 1). Dimensional constraints are applied by setting a "formula". 3) In the "Assembly Design" workbench interface, match the partition plate instances with dot1; use the "Reuse Array" command to perform an instance array of partition plate instances based on the rectangular array in the previous step, and "keep the link with the array". Thus, changing the "number of partition plates" controls the number of partition plate instances in the main beam.

[0062] Exporting parameters as a design table: Using CATIA's "Design Table" function, create a design table in *.txt format using the current parameter values, and associate the parameterized variables with the parameters in the design table. After modifying the parameters in the txt file, updating the CATIA software will synchronously modify the geometric model.

[0063] A catvbs script macro is written to update the parameters of each component of the main beam after modification. Finally, the geometric model is output in STP format, a file supported by the finite element preprocessing software. The update script is: `set product1 = productdocument1.product; product1.update`. The geometric output script is: `set partdocument1 = catia.activedocument; partdocument1.exportdata`. The file is saved as `.stp","stp"`.

[0064] Step (2): Import the geometric file into the finite element preprocessing software;

[0065] Import the STP geometry file into Hypermesh software. In CATIA software, use the "Measurement Items" command to calculate the area A1 of the mid-surface geometry; in Hypermesh software, use the "mass calc" command to calculate the area A2 of the mid-surface. Determine if A1 equals A2 to verify the correctness of the geometric model's process from design software output to preprocessing software input, and check for errors such as missing or damaged surfaces. If A1 ≠ A2, adjust the 3D geometry and STP file export parameters until the requirements are met.

[0066] Step (3): Preprocessing for finite element simulation analysis of the main beam structure;

[0067] The model components are connected according to welding, bolting, and other relationships to simulate actual welded connections. Intersecting surfaces such as the upper and lower flanges and web, the diaphragm and the upper flange and web, the stiffener and web, and the end plate and the upper and lower flanges and web are connected using surface extension to achieve shared mesh nodes along the intersection lines. The upper and lower contact surfaces of the lower flange and the end beam connecting plate are treated by superimposing surface thicknesses to merge geometric surfaces. Utilizing the relatively fixed ID numbers and component names of the imported geometric model elements, automatic processing by the script program is possible. Geometric subdivision is performed on the web, upper and lower flanges, and diaphragm intersection surfaces based on the ID numbers of geometric points, lines, and surfaces, followed by mesh generation and smoothing.

[0068] The main beam components are treated with shell elements, while the trolley tracks are treated with solid elements. To improve mesh adaptability, a combination of quadrilateral and triangular meshing strategies is used for the shell elements. The mesh quality is checked according to the element quality requirements in the "General Rules for Finite Element Mechanical Analysis of Mechanical Product Structures". Quadrilateral elements have an aspect ratio ≤ 5.0, warpage ≤ 16°, skewness ≤ 60°, and interior angles of 40° to 135°; triangular elements have an aspect ratio ≤ 5.0, skewness ≤ 60°, and interior angles of 20° to 120°.

[0069] For meshes that do not meet the requirements, perform surface subdivision and geometric feature processing until the mesh quality meets the requirements.

[0070] The material and section properties are assigned to the elastoplastic body, and the thickness information is assigned to the section properties of the plate and shell element.

[0071] The model was weight-checked. In the HyperMesh software, the "mass calc" command was used to calculate the total weight M1 of all elements in the finite element model and compare it with the designed total weight M2 of the main beam to check the correctness of the model's mesh, material properties, and section property settings. If M1 ≠ M2, the mesh, material data, and section property settings were checked and corrected.

[0072] Load settings are performed according to the national standard "Crane Design Code" and design requirements, and load steps, load combinations, and boundary constraints are established. The wheel pressure of the trolley is coupled to the corresponding position on the track in the form of concentrated force.

[0073] Step (4): The pre-processing process of the main beam structure is further developed to form a standardized processing script program;

[0074] The preprocessing in step (3) uses script commands to form a standardized TCL language processing flow program. The plate thickness of each component of the main beam is set in the script to facilitate subsequent parameterization. The script program automatically imports the geometric model file, performs connection, mesh, load, constraint, and load step processing, and outputs an Abaqus solution file in .inp format.

[0075] To achieve mesh node sharing through face extension, the *connect_surfaces_11 function is used, as follows: *createmarksurfaces 1 face ID number…; *createmark lines 1; *createmark lines 2; *connect_surfaces_11 1 1 1 2 3 15 30 1 1 2 30 3 0

[0076] Face segmentation uses the *surfmark_trim_by_surfmark function, as follows: `set names_list1[list name of component to be segmented 1 name of component to be segmented 2 ...]; eval *createmark surfaces 1 "by collector name" $names_list1; set names_list2[list name of segmented component 1 name of segmented component 2 ...]; eval *createmark surfaces 2 "by collector name" $names_list2; *surfmark_trim_by_surfmark 1 2 2`

[0077] Mesh generation is performed using the *automesh function, and the *set_meshfaceparams function is used to set the mesh parameters first.

[0078] Smoothing the mesh uses the *marksmoothelements function, as follows: *clearmark elements 1; *createmark elements 1"displayed"; *createmark nodes 1 *marksmoothelements 1 1 110

[0079] Section properties are created using the `*createentity props` method: `*createentity props cardimage = SHELLSECTION name = $propName; set mat_id [hm_getvalue mats name = $matName dataname = id]; *setvalue props name = $propName materialid = {mats$mat_id}; *setvalue props name = $propName STATUS = 1111 = $thick`. Section property settings are: `*setvalue comps mark = 1 propertyid = {props$propid}`.

[0080] Load steps are created using *createentity loadsteps name = loadstep_1. Abaqus is used as the solver. Load settings are configured using the *setvalue method, such as: loadsteps id = $loadstep_id STATUS = 2 195 = 1.

[0081] Inp file output method: set hypermesh_path[hm_info-appinfo altair_home]; append hypermesh_path" / templates / feoutput / abaqus / standard.3d"; *feoutputwithdata$hypermesh_path$inp_file_path 0 0 2 1 2

[0082] Step (5): Calculate using the finite element solution file and output the calculation results for strength, stiffness, stability, etc.

[0083] Import the inp solution file into the Abaqus finite element software solver for calculation, and extract results such as strength (Mises stress, principal stress), stiffness (displacement), and stability (buckling eigenvalue) from the result file, while also outputting the model's weight information.

[0084] Step (6): Based on finite element DOE analysis, establish an approximate proxy model for the simulation model;

[0085] A Design of Effect (DOE) workflow was established using the iSight software. Optimized Latin squares were used to sample the aforementioned key dimensions and thickness parameters within the design space. Let the number of parameters be M, and the number of sample data points be N, then N > (M+1)*(M+2) / 2. To improve calculation accuracy, N was set to 200 in this example. The Catia geometric model parameterization update used the Simcode component to read and modify the txt parameter table file, and then called a batch command to execute a catvbs script to update and output the STP geometric model. Hypermesh finite element preprocessing used the Simcode component to call a batch command to execute a tcl script, outputting an .inp format Abaqus finite element solution file. The batch command was hmbatch.exe "-tcl"preprocessingscript.tcl". Solving the .inp file and extracting results used the iSight Abaqus component. The Abaqus component read the .inp file, parameterized the plate thickness of each component, solved the .inp file, and extracted stress, displacement, buckling characteristic values, and weight information to form sample point data.

[0086] Based on the design parameters and the response data from the calculations, an approximate surrogate model is established using the RBF neural network algorithm. The fitting parameters are: Smoothing filter = 0.02, Type of Bass Function = Elliptical, Maximum Iterations to Fit = 100. Error analysis employs cross-validation, with the error analysis sample representing 20% ​​of the total sample size (40 in this example). The goodness-of-fit R-squared of the surrogate model is then evaluated. 2 Error analysis and evaluation, such as R 2 If the value is greater than 0.9, proceed to the next step; otherwise, adjust the model fitting parameters or add sample data to refit until the goodness-of-fit requirement is met, at which point the approximate surrogate model can be used.

[0087] Step (7): Optimize the main beam of the crane based on the approximate surrogate model that meets the requirements.

[0088] In the Optimization module of the Isight software, the parameterized dimensions from the aforementioned steps are used as design variables. Based on the actual manufacturing constraints of the height and width of the main beam section and the variation range of the plate thickness of each component in the design space, and with constraints such as strength (maximum stress) <140MPa, stiffness (maximum downward deflection under static stiffness conditions) <26mm, and stability (buckling characteristic value) > design requirements, the optimization objective is to minimize the weight of the main beam. The Multi-Island Genetic Algorithm (Multi-IslandGA) is used to iteratively optimize based on the surrogate model.

[0089] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0090] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A parametric modeling and optimization design method for the box girder of a gantry crane, characterized in that, include: Parametric geometric modeling of the crane's main beam was implemented in 3D modeling software, and the geometric model was output. The geometric model output by the 3D design software is automatically imported into the finite element preprocessing software, and the geometric model is preprocessed to establish a parametric finite element simulation model of the crane main beam. Preprocessing is performed on the parametric finite element simulation model of the main beam of the crane for finite element simulation analysis, and the preprocessed parametric finite element simulation model is solved and calculated. An approximate proxy model for the parametric finite element simulation model of the crane main beam is established based on DOE analysis of finite element method. Based on the parametric finite element simulation model of the crane main beam, optimization algorithms are used to find the best model under safety and manufacturability constraints for the approximate surrogate model constructed by DOE analysis. The preprocessing for the parametric finite element simulation analysis of the crane's main beam structure includes: The components of the parametric finite element simulation model are connected according to welding and bolting relationships to simulate actual weld connections. Utilizing the relatively fixed ID numbers and component names of the imported geometric model elements, geometric subdivision is performed on the intersection surfaces of the web, upper and lower flanges, and diaphragms based on the ID numbers of geometric points, lines, and surfaces. Mesh generation and smoothing are then performed. The main beam components use shell elements, while the trolley track uses solid elements. Material and section properties are assigned to elastoplastic bodies, and loads are set, including load steps, load combinations, and boundary constraints. The parametric finite element simulation model undergoes weight verification. The total weight of the parametric finite element simulation model is calculated and compared with the total design weight of the main beam to check the correctness of the mesh, material properties, and section property settings. If the total weight of the parametric finite element simulation model is not equal to the total design weight of the main beam, the mesh, material data, and section property settings are checked and corrected.

2. The method according to claim 1, characterized in that, The parametric geometric modeling of the crane's main beam in 3D modeling software, and the output of the geometric model, includes: The geometric models of the upper and lower flange plates, web plates, diaphragm plates, end plates, and stiffening ribs of the box girder of the gantry crane were established using 3D modeling software. A solid 3D model of the track was also established. The geometric model plates were modeled using the mid-surface, and geometric features with little impact on the calculation were ignored. The width of the upper and lower flange plates, the height of the web plates, the spacing between the web plates, the number of intermediate diaphragm plates, and the spacing of the stiffening ribs of the main girder were parameterized to form a parameter table. The geometric model was output in a file format supported by the finite element preprocessing software.

3. The method according to claim 2, characterized in that, The automatic import of the geometric model output by the 3D design software into the finite element preprocessing software, and the preprocessing of the geometric model, include: The area of ​​the mid-surface geometry is calculated in the 3D modeling software and compared with the area of ​​the mid-surface geometry calculated in the finite element software to determine the correctness of the geometric model output from the 3D modeling software to the input of the finite element preprocessing software. This is used to perform preprocessing on the geometric model until the geometric model in the finite element preprocessing software is consistent with the geometric model in the 3D modeling software, thus obtaining the parametric finite element simulation model of the crane main beam.

4. The method according to claim 3, characterized in that, The process of solving the preprocessed parametric finite element simulation model includes: The pre-processing process of the main beam structure is further developed to form a standardized processing script program. The plate thickness of each component of the main beam is set in the script program. The script program automatically imports geometric model files, performs connection, mesh, load, constraint, and load step processing, and outputs finite element solution files. Import the solution file into the finite element software solver for calculation, and extract the strength, stiffness and stability results from the result file to output the weight of the parameterized finite element simulation model.

5. The method according to claim 4, characterized in that, The approximate proxy model for establishing the parametric finite element simulation model of the crane main beam based on the finite element DOE analysis includes: For key parameters such as dimensions and thickness of components within the design space, sampling is performed according to the DOE experimental design method. The geometric model is automatically modified based on the changing parameters. A standardized preprocessing script is used to quickly generate a parametric finite element simulation model and output the solution file. The finite element solver is called to calculate the output results. Strength, stiffness, stability results and weight information are extracted to form sample point data. Based on the design parameter variables and the response data of the calculation results, an approximate surrogate model is established using a model fitting algorithm. The goodness-of-fit R² error analysis of the surrogate model is performed to evaluate the surrogate model and obtain an approximate surrogate model that meets the requirements.

6. The method according to claim 5, characterized in that, The error analysis and evaluation of the goodness-of-fit R² of the surrogate model to obtain a satisfactory approximate surrogate model includes: The surrogate model is evaluated by error analysis using the goodness-of-fit R². If R² > 0.9, proceed to the next step; otherwise, adjust the fitting parameters of the surrogate model, change the fitting algorithm, or add sample data to refit until the goodness-of-fit requirement is met, thus obtaining a satisfactory approximate surrogate model.

7. The method according to claim 6, characterized in that, Based on the parametric finite element simulation model of the crane main beam, the optimization algorithm is used to perform optimization under safety and manufacturability constraints on the approximate surrogate model constructed by DOE analysis, including: Using parametric dimensions as design variables, the design space is limited by actual production and manufacturing conditions, with strength, stiffness, stability and manufacturability indicators as constraints, and the weight of the main beam as the optimization objective. An intelligent optimization algorithm is adopted to iteratively optimize based on a surrogate model.

Images