Hull structure finite element mesh optimization method and system
By extracting and utilizing the key feature data of the finite element mesh model and combining artificial intelligence algorithm to train the optimization model, the problems of inefficiency and inconsistent results in the optimization of the finite element mesh in the hull structure are solved, achieving more efficient and consistent optimization effects.
Patent Information
- Application Number
- CN202510235614.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-28
- Publication Date
- 2025-04-18
AI Technical Summary
The prior art relies on manual repair and optimization in the optimization of finite element grids of hull structures, resulting in inefficiency and inconsistent results, and failure to effectively utilize design experience in historical optimization cases.
By extracting the key feature data of the finite element mesh model, building a training data set, and using artificial intelligence algorithms to train the grid optimization algorithm model, automatic optimization of the target finite element mesh model is achieved.
It significantly improves the efficiency and quality consistency of finite element grid optimization, eliminates artificial experience differences, and ensures the standardization of grid adjustment strategies with the same or similar structural characteristics.
Smart Images
Figure CN119849259A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of ship design, and in particular to a finite element mesh optimization method and system for a hull structure. Background Art
[0002] Due to the complex structure and large number of parts, the finite element mesh generation of the hull structure is usually processed in stages. The existing technology generally performs preliminary automatic meshing based on the hull design model to generate the structural finite element mesh, and then the designer manually repairs and optimizes the mesh model until the calculation accuracy requirements are met. This traditional method has significant defects: first, the manual repair process involves a large number of repetitive operations, and the work efficiency is low; second, the mesh quality is highly dependent on the individual experience accumulation of the designer, and it is difficult to ensure the consistency of the optimization results.
[0003] It is worth noting that shipbuilding enterprises have accumulated a large number of structural finite element mesh model databases that have been verified by manual optimization in long-term practice. These data essentially contain the cognitive experience of designers on complex structural mesh optimization rules. However, the existing technology has not yet established an effective knowledge mining mechanism to explicitly represent these implicit expert experiences, nor has it built a systematic rule extraction method to transform historical optimization results into reusable automated processing strategies, resulting in low utilization of knowledge assets and restricting the further improvement of mesh optimization efficiency. Summary of the invention
[0004] The purpose of this application is to provide a finite element mesh optimization method and system for hull structure, which transforms implicit experience in historical optimization cases into explicit algorithm models through artificial intelligence training, realizes the systematic inheritance and reuse of design knowledge, improves optimization efficiency and quality consistency, and solves the problems of strong manual dependence, low efficiency and insufficient utilization of historical data.
[0005] In a first aspect, a finite element mesh optimization method for a hull structure is provided, comprising:
[0006] S1. preparing a plurality of finite element mesh models and extracting key characteristic data of the finite element mesh models; the plurality of finite element mesh models include a finite element mesh model before optimization and a corresponding finite element mesh model after optimization;
[0007] S2, using the key feature data of the finite element mesh model before optimization as input data and the key feature data of the finite element mesh model after optimization as output data to construct a finite element mesh training data set;
[0008] S3, select a preset artificial intelligence algorithm, and input a finite element mesh training data set to obtain a mesh optimization algorithm model through training;
[0009] S4, obtaining key feature data of the target finite element mesh model to be optimized, inputting the mesh optimization algorithm model, and obtaining corresponding output data as the optimized key feature data;
[0010] S5. Adjust the target finite element mesh model to be optimized based on the optimized key feature data to obtain the final optimized finite element mesh model.
[0011] In an practicable solution, after step S5, the following further steps are included:
[0012] S6. Use the key feature data to be optimized of the target finite element mesh model as input data, use the optimized key feature data as output data, add them to the finite element mesh training data set, and repeat step S3.
[0013] In an practicable solution, before step S3, the following steps are further included:
[0014] Establish an artificial intelligence algorithm library; the artificial intelligence algorithm library stores a number of preset artificial intelligence algorithms, and different preset artificial intelligence algorithms are suitable for the training of grid optimization algorithm models of different ship types and different compartments;
[0015] In step S3, the step of selecting a preset artificial intelligence algorithm includes:
[0016] Determine the ship type and compartment to which the structure belongs;
[0017] Determine and select applicable preset AI algorithms based on ship type and compartment.
[0018] In an practicable solution, step S3 further includes:
[0019] Establish an optimization algorithm model library, store the trained grid optimization algorithm model, and bind it with the ship type and compartment data.
[0020] In an practicable solution, in step S4, after the step of obtaining the key feature data of the target finite element mesh model to be optimized and before the step of inputting the mesh optimization algorithm model, the method further includes:
[0021] Determine the ship type and cabin section to which the target finite element mesh model belongs;
[0022] Based on the ship type and compartment data, an applicable grid optimization algorithm model is selected from the optimization algorithm model library.
[0023] In one feasible solution, in step S1, the step of preparing a plurality of finite element mesh models includes:
[0024] Determine a finite element mesh quality evaluation standard; the finite element mesh quality evaluation standard is used to determine that the current finite element mesh model is an optimized finite element mesh model;
[0025] The optimized finite element mesh model is determined based on the finite element mesh quality evaluation standard, and the corresponding finite element mesh model before optimization is found and stored in a mesh model library in a paired binding form.
[0026] In an practicable solution, in step S1, the step of extracting key feature data of the finite element mesh model includes:
[0027] Extracting key feature data of each finite element mesh model in the mesh model library in sequence and storing them in a training database;
[0028] Among them, key feature data of the optimized finite element mesh model and the finite element mesh model before optimization are set with an associated relationship in the training database.
[0029] In one feasible solution, the key feature data includes label data and geometric information;
[0030] The tag data at least includes ship type data and compartment type data;
[0031] The geometric information at least includes mesh node information and mesh unit information;
[0032] The grid node information includes the node number and the node coordinates of each grid node;
[0033] The grid cell information includes the cell number, the cell type, and the node numbers of the included grid nodes.
[0034] In a second aspect, a finite element mesh optimization system for a hull structure is also provided, comprising:
[0035] Mesh model management module, used to store, delete and add finite element mesh models;
[0036] Data extraction module, used to extract key feature data of finite element mesh model;
[0037] A data management module, used to construct a finite element mesh training data set using key feature data of the finite element mesh model before optimization as input data and key feature data of the finite element mesh model after optimization as output data;
[0038] The algorithm model training module is used to select a preset artificial intelligence algorithm and input a finite element mesh training data set to train and obtain a mesh optimization algorithm model;
[0039] The data optimization module is used to obtain the key feature data to be optimized of the target finite element mesh model, and input the corresponding mesh optimization algorithm model to obtain the corresponding output data as the optimized key feature data;
[0040] The optimization output module is used to adjust the target finite element mesh model to be optimized based on the optimized key feature data and output the final optimized finite element mesh model.
[0041] In an implementable scheme, the hull structure finite element mesh optimization system also includes a data set update module; the data set update module is used to use the key feature data of the target finite element mesh model obtained by the data optimization module as input data, and use the optimized key feature data as output data to supplement the finite element mesh training data set.
[0042] Compared with the prior art, the beneficial effects of this application include at least:
[0043] The finite element mesh optimization method of the hull structure of this application realizes the systematic inheritance and reuse of design knowledge by converting implicit experience in historical optimization cases into explicit algorithm models, thereby significantly improving optimization efficiency and quality consistency.
[0044] The method of this application extracts key feature data, uses artificial intelligence algorithms to learn optimization rules from historical cases, and encodes the designer's implicit decision logic into a reusable algorithm model. This process eliminates differences in human experience and ensures the standardization of grid adjustment strategies for the same or similar structural features.
[0045] At the same time, in this application, after the key feature data of the model to be optimized is input into the trained algorithm model, the system directly generates the optimization parameters and automatically adjusts the grid, avoiding the tedious operation of manual point-by-point correction.
[0046] Through structured feature extraction and mapping mechanisms, discrete optimization decisions are transformed into trainable data relationships, making massive historical cases a source of knowledge for algorithm iteration. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0048] Figure 1 It is a flow chart of a finite element mesh optimization method for a hull structure according to an embodiment of the present application;
[0049] Figure 2 It is a structural schematic diagram of a finite element mesh optimization system for a hull structure according to an embodiment of the present application;
[0050] Figure 3 It is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION
[0051] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.
[0052] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.
[0053] like Figure 1 As shown, the embodiment of the present application first provides a finite element mesh optimization method for a hull structure, comprising:
[0054] S1. preparing a plurality of finite element mesh models and extracting key characteristic data of the finite element mesh models; the plurality of finite element mesh models include a finite element mesh model before optimization and a corresponding finite element mesh model after optimization;
[0055] S2, using the key feature data of the finite element mesh model before optimization as input data and the key feature data of the finite element mesh model after optimization as output data to construct a finite element mesh training data set;
[0056] S3, select a preset artificial intelligence algorithm, and input a finite element mesh training data set to obtain a mesh optimization algorithm model through training;
[0057] S4, obtaining key feature data of the target finite element mesh model to be optimized, inputting the mesh optimization algorithm model, and obtaining corresponding output data as the optimized key feature data;
[0058] S5. Adjust the target finite element mesh model to be optimized based on the optimized key feature data to obtain the final optimized finite element mesh model.
[0059] This application proposes an artificial intelligence-based finite element mesh optimization method for hull structures to solve the problems of strong manual dependence, low efficiency and insufficient utilization of historical data in finite element mesh optimization of hull structures. By converting implicit experience in historical optimization cases into explicit algorithm models, the systematic inheritance and reuse of design knowledge is achieved, significantly improving optimization efficiency and quality consistency.
[0060] The traditional method relies on human experience, which leads to inconsistent optimization results. This application solves the problem by constructing a training data set for model comparison before and after optimization. This method extracts key feature data, uses artificial intelligence algorithms to learn optimization rules from historical cases, and encodes the designer's implicit decision logic into a reusable algorithm model. This process eliminates differences in human experience and ensures the standardization of mesh adjustment strategies for the same or similar structural features.
[0061] To address the problem of inefficient repetitive operations, the solution speeds up the optimization process through automated parameter adjustment. After the key feature data of the optimization model is input into the trained algorithm model, the system directly generates the optimization parameters and automatically adjusts the grid, avoiding the tedious operation of manual point-by-point correction. The pain point of insufficient utilization of historical data is solved through the structured feature extraction and mapping mechanism, which converts discrete optimization decisions into trainable data relationships, making massive historical cases a source of knowledge for algorithm iteration.
[0062] In one embodiment, the key feature data may include label data and geometric information. The label data may include ship type data and compartment type data. The geometric information may include grid node information and grid unit information, and the grid node information may include node number and node coordinates of each grid node. The grid unit information may include unit number, unit type and node number of the included grid nodes. As shown in Table 1 and Table 2, they are key feature data tables before and after optimization of a finite element grid model, respectively.
[0063] Table 1 Schematic diagram of key characteristic data of finite element mesh model before optimization
[0064]
[0065]
[0066] Table 2 Schematic diagram of key characteristic data of optimized finite element mesh model
[0067]
[0068] In one embodiment, in step S1, after extracting key feature data of the finite element mesh model, the data is preprocessed.
[0069] For a certain hull structure, the essential difference between its finite element mesh model before and after optimization lies in the change in the mesh composition method. The finite element mesh model before optimization usually does not meet the calculation requirements, such as the existence of node overlap, too many small units and other problems, while the finite element mesh model after optimization adjusts the mesh composition method, such as adjusting the node position, changing the unit type, etc., so that the structural finite element mesh model can meet the calculation requirements.
[0070] Among them, preprocessing can be to clean the key feature data of the extracted finite element mesh model automatically or manually, including: completing, correcting or deleting label data and geometric information to ensure the accuracy of the data. For example, for mesh node information, it is necessary to check whether the node coordinates of each node number have 3 values, that is, the x, y, and z coordinate values of the node, and whether the coordinate values are within the hull coordinate system; for mesh unit information, it is necessary to check whether the quadrilateral unit has 4 node numbers, whether the triangular unit has 3 node numbers, whether the one-dimensional unit has 2 node numbers, and whether each node number can be found in the node information, and correct the correct or erroneous feature data.
[0071] In one embodiment, after step S5, the method further includes:
[0072] S6. Use the key feature data to be optimized of the target finite element mesh model as input data, use the optimized key feature data as output data, add them to the finite element mesh training data set, and repeat step S3.
[0073] In step S6, a data update threshold (for example, set to 50) and a data update variable can be set. Before repeating step S3, the data update variable is automatically increased by one for each additional set of input data and output data in the finite element mesh training data set. And monitor at any time whether the value of the data update variable is greater than or equal to 50; if so, repeat step S3; if not, keep the status quo and do not repeat step S3.
[0074] After step S6 is added, the model can be continuously upgraded by continuously absorbing new optimization cases. This method not only increases the optimization speed several times, but more importantly, establishes a dedicated grid optimization knowledge base. Subsequent new users (new employees) can quickly master the optimization skills of senior designers with the help of the system associated with this method, and the successful experience in historical projects can also be systematically reused in new ship design.
[0075] In one embodiment, before step S3, the step further includes:
[0076] Establishing an artificial intelligence algorithm library; the artificial intelligence algorithm library stores a number of preset artificial intelligence algorithms, and different preset artificial intelligence algorithms are suitable for training grid optimization algorithm models of different ship types and different compartments;
[0077] In step S3, the step of selecting a preset artificial intelligence algorithm includes:
[0078] Determine the ship type and compartment to which the structure belongs;
[0079] Determine and select applicable preset AI algorithms based on ship type and compartment.
[0080] In one embodiment, the step S3 further includes:
[0081] Establish an optimization algorithm model library, store the trained grid optimization algorithm model, and bind it with the ship type and compartment data.
[0082] This is equivalent to conducting separate training based on different types of ship types, cabin sections, etc., to obtain multiple algorithm models suitable for optimizing different finite element mesh models, ensuring the accuracy of subsequent intelligent optimization models.
[0083] For example, when training the cargo hold and engine room areas of a ship, the structures of the two areas are quite different, that is, the cargo hold area has a simple structure, many straight areas, a large grid size and a low density; while the engine room area has a complex structure, few straight areas, a small grid size and a high density. The above structural differences require the use of different training data for the two, that is, the training data for the cargo hold area is mainly based on the grid data of the cargo hold area, and the training data for the engine room area is mainly based on the grid data of the engine room area. In terms of the selection of training algorithms, supervised learning algorithms such as neural networks and random forests can be selected for the training of the two areas. For example, by building a deep neural network to match the relationship between the grid data before and after optimization, automatic optimization can be achieved. If the training effect of a single algorithm is not good, a combination of multiple algorithms can be used for training, such as a combination of neural networks and random forests. Taking the neural network algorithm as an example, the main steps of the algorithm include:
[0084] (1) Network architecture design, determining the number of network layers, number of neurons, and activation functions;
[0085] (2) Submit the key feature data before optimization as input data to the network, and calculate the output results based on the network structure;
[0086] (3) Calculate the loss function between the output result and the optimized key feature data to determine the difference between the two;
[0087] (4) Based on the gap calculated in the previous step, the network parameters are optimized. Through multiple calculations and parameter iterative optimization, it is ensured that the network can accurately describe the relationship between the grid before and after optimization.
[0088] In one embodiment, in step S4, after the step of obtaining key feature data of the target finite element mesh model to be optimized and before the step of inputting the mesh optimization algorithm model, the step further includes:
[0089] Determine the ship type and cabin section to which the target finite element mesh model belongs;
[0090] Based on the data of ship type and compartment, the applicable grid optimization algorithm model is selected from the optimization algorithm model library to ensure the matching of the model and improve the accuracy of the prediction.
[0091] In one embodiment, in step S1, the step of preparing a plurality of finite element mesh models includes:
[0092] Determining a finite element mesh quality evaluation standard; the finite element mesh quality evaluation standard is used to determine that the current finite element mesh model is an optimized finite element mesh model;
[0093] The optimized finite element mesh model is determined based on the finite element mesh quality evaluation standard, and the corresponding finite element mesh model before optimization is found and stored in a mesh model library in a paired binding form.
[0094] Among them, the finite element mesh quality evaluation standard can be comprehensively judged through quantitative dimensions such as unit shape quality parameters, size adaptability parameters, connection standardization parameters and boundary fit parameters.
[0095] In one embodiment, in step S1, the step of extracting key feature data of the finite element mesh model includes:
[0096] Extracting key feature data of each finite element mesh model in the mesh model library in sequence and storing them in a training database;
[0097] Among them, key feature data of the optimized finite element mesh model and the finite element mesh model before optimization are set with an associated relationship in the training database.
[0098] like Figure 2 As shown, the embodiment of the present application also provides a finite element mesh optimization system for a hull structure, comprising:
[0099] A mesh model management module 10, used for storing, deleting and adding finite element mesh models;
[0100] A data extraction module 20, used to extract key feature data of the finite element mesh model;
[0101] A data management module 30, used to construct a finite element mesh training data set using the key feature data of the finite element mesh model before optimization as input data and the key feature data of the finite element mesh model after optimization as output data;
[0102] The algorithm model training module 40 is used to select a preset artificial intelligence algorithm and input a finite element mesh training data set to train and obtain a mesh optimization algorithm model;
[0103] The data optimization module 50 is used to obtain the key feature data to be optimized of the target finite element mesh model, and input the corresponding mesh optimization algorithm model to obtain the corresponding output data as the optimized key feature data;
[0104] The optimization output module 60 is used to adjust the target finite element mesh model to be optimized based on the optimized key feature data, and output the final optimized finite element mesh model.
[0105] In one embodiment, if Figure 2 As shown, the finite element mesh optimization system for hull structure also includes a data set updating module 70; the data set updating module is used to use the key feature data of the target finite element mesh model obtained by the data optimization module as input data, and use the optimized key feature data as output data to supplement the finite element mesh training data set.
[0106] Corresponding to the above method embodiment, Figure 3 As shown, the present application also provides an electronic device, which includes a processor 1 and a memory 2 storing computer program instructions. The processor 1 and the memory 2 can communicate with each other through a communication bus 3. The processor 1 is used to execute computer program instructions, and the memory 2 is used to store computer program instructions.
[0107] The processor 1 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the present embodiment. The memory 2 may include a high-speed RAM memory, or may also include a non-volatile memory, such as at least one disk memory. The processor 1 executes at least one computer program instruction to implement the steps shown in the embodiment of the finite element mesh optimization method for the hull structure.
[0108] Corresponding to the above method embodiments, the present application also provides a computer-readable storage medium, on which computer program instructions are stored. When one or more computer program instructions are executed by a processor, the steps shown in the embodiment of the finite element mesh optimization method for a hull structure are implemented.
[0109] Among them, computer storage media can be tangible media that can contain or store programs for use by or in combination with instruction execution systems, devices or equipment. Computer storage media can be machine-readable signal media or machine-readable storage media. Computer storage media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.
[0110] Corresponding to the above method embodiment, the present application also provides a computer program product, and when the instructions in the computer program product are executed by a processor, the steps shown in the embodiment of the finite element mesh optimization method for a hull structure are implemented.
[0111] In summary, the technical solution of this application uses a machine learning algorithm to automatically learn the rules and experience of mesh optimization from a large number of accumulated finite element mesh models of hull structures. When processing new finite element mesh models of hull structures, the mesh is automatically optimized and adjusted based on the learned rules and experience, reducing manual processing work. Compared with relying entirely on manual mesh optimization work, this method can improve efficiency, avoid errors that occur during manual processing, and achieve intelligent optimization of finite element meshes of hull structures.
[0112] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A finite element mesh optimization method for a hull structure, characterized in that: include: S1. Prepare several finite element mesh models and extract key feature data of the finite element mesh models; The plurality of finite element mesh models include a finite element mesh model before optimization and a corresponding finite element mesh model after optimization; S2, using the key feature data of the finite element mesh model before optimization as input data and the key feature data of the finite element mesh model after optimization as output data to construct a finite element mesh training data set; S3, select a preset artificial intelligence algorithm, and input a finite element mesh training data set to obtain a mesh optimization algorithm model through training; S4, obtaining key feature data of the target finite element mesh model to be optimized, inputting the mesh optimization algorithm model, and obtaining corresponding output data as the optimized key feature data; S5. Adjust the target finite element mesh model to be optimized based on the optimized key feature data to obtain the final optimized finite element mesh model.
2. The finite element mesh optimization method for hull structure according to claim 1, characterized in that: After step S5, the method further includes: S6. Use the key feature data to be optimized of the target finite element mesh model as input data, use the optimized key feature data as output data, add them to the finite element mesh training data set, and repeat step S3.
3. The finite element mesh optimization method for hull structure according to claim 1, characterized in that: Before step S3, the following steps are also included: Establishing an artificial intelligence algorithm library; the artificial intelligence algorithm library stores a number of preset artificial intelligence algorithms, and different preset artificial intelligence algorithms are suitable for training grid optimization algorithm models of different ship types and different compartments; In step S3, the step of selecting a preset artificial intelligence algorithm includes: Determine the ship type and section to which the structure belongs; Determine and select applicable preset AI algorithms based on ship type and compartment.
4. The finite element mesh optimization method for hull structure according to claim 1, characterized in that: The step S3 further comprises: Establish an optimization algorithm model library, store the trained grid optimization algorithm model, and bind it with the ship type and compartment data.
5. The finite element mesh optimization method for hull structure according to claim 4, characterized in that: In the step S4, after the step of obtaining the key feature data of the target finite element mesh model to be optimized and before the step of inputting the mesh optimization algorithm model, the step further includes: Determine the ship type and cabin section to which the target finite element mesh model belongs; Based on the ship type and compartment data, an applicable grid optimization algorithm model is selected from the optimization algorithm model library.
6. The finite element mesh optimization method for hull structure according to claim 1, characterized in that: In step S1, the step of preparing a plurality of finite element mesh models includes: Determining a finite element mesh quality evaluation standard; the finite element mesh quality evaluation standard is used to determine that the current finite element mesh model is an optimized finite element mesh model; The optimized finite element mesh model is determined based on the finite element mesh quality evaluation standard, and the corresponding finite element mesh model before optimization is found and stored in a mesh model library in a paired binding form.
7. The finite element mesh optimization method for a hull structure according to claim 6, characterized in that: In step S1, the step of extracting key feature data of the finite element mesh model includes: Extracting key feature data of each finite element mesh model in the mesh model library in sequence and storing them in a training database; Among them, key feature data of the optimized finite element mesh model and the finite element mesh model before optimization are set with an associated relationship in the training database.
8. The finite element mesh optimization method for a hull structure according to any one of claims 1 to 7, characterized in that: The key feature data includes label data and geometric information; The tag data at least includes ship type data and compartment type data; The geometric information includes at least grid node information and grid unit information; The grid node information includes a node number and a node coordinate of each grid node; The grid unit information includes a unit number, a unit type, and node numbers of included grid nodes.
9. A finite element mesh optimization system for a hull structure, characterized in that: include: Mesh model management module, used to store, delete and add finite element mesh models; Data extraction module, used to extract key feature data of finite element mesh model; A data management module, used to construct a finite element mesh training data set using key feature data of the finite element mesh model before optimization as input data and key feature data of the finite element mesh model after optimization as output data; The algorithm model training module is used to select a preset artificial intelligence algorithm and input a finite element mesh training data set to train and obtain a mesh optimization algorithm model; The data optimization module is used to obtain the key feature data to be optimized of the target finite element mesh model, and input the corresponding mesh optimization algorithm model to obtain the corresponding output data as the optimized key feature data; The optimization output module is used to adjust the target finite element mesh model to be optimized based on the optimized key feature data and output the final optimized finite element mesh model.
10. The finite element mesh optimization system for hull structure according to claim 9, characterized in that: It also includes a data set updating module; the data set updating module is used to use the key feature data of the target finite element mesh model obtained by the data optimization module as input data, and use the optimized key feature data as output data to supplement the finite element mesh training data set.