Ship body section shelving deformation simulation analysis method considering time effect

Abstract

The invention provides a ship body section shelving deformation simulation analysis method considering a time effect, and aims to solve the safety problem that an evaluation result deviates from reality due to the fact that a time factor is not counted in a long-term shelving scene in traditional finite element analysis. According to the method, stress and deformation data of existing real ship construction segments in shelving states in different time periods are obtained, a segment structure stress and deformation data set in an ideal state is obtained through a segment structure simulation model, and a structure simulation analysis deduction model under a time effect is constructed in combination with a machine learning algorithm. Therefore, analysis of structural deformation of a time factor is added. According to the method, the stress and deformation data of the segmented shelving state under the time effect are obtained through a machine learning algorithm, the stress and deformation under the time effect are further considered on the basis of traditional finite element analysis, the segmented shelving safety is evaluated more comprehensively, a judgment basis is provided for a shelving scheme, and the safety of the shelving scheme is improved. And the safety and reliability of the ship building process are ensured.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding technology, and in particular to a simulation analysis method for deformation of ship sections under suspension that takes into account time effects. Background Technology

[0002] Shipbuilding consists of intermediate products such as sections, main sections, and mega-sections. After completion, ship sections are placed on gantry cranes, piers, and other tooling for a period before being assembled into main sections. During this placement, sections undergo structural deformation, necessitating structural simulation analysis to determine deformation values ​​and the installation of appropriate support fixtures. However, this deformation changes over time, and traditional finite element methods can only consider initial deformation under theoretical conditions, failing to account for structural deformation over time. Consequently, the design of support fixtures cannot fully account for the influence of time. Summary of the Invention

[0003] In view of the problems existing in the structural simulation of the segmented and shelved state, the present invention provides a structural simulation analysis method that considers the time effect. By obtaining stress and deformation data of existing ship construction segments under different time periods of shelving, and combining machine learning algorithms to construct a structural simulation analysis and deduction model under the time effect, the analysis of structural deformation with the time factor is realized.

[0004] To achieve the above and other related objectives, this invention provides a simulation analysis method for the deformation of ship hull sections under suspension, considering time effects, comprising the following steps:

[0005] S1. Constructing structural simulation analysis and deduction models under time effects for typical structural sections of different ship types; specifically including:

[0006] S11. Identify typical structural sections for different ship types;

[0007] S12. Select the existing segments as feature segments, select key locations of the segments as sampling points, arrange strain sensors at the sampling points, uniformly set strain acquisition time points during the segment rest time interval, collect strain data of each strain sensor at regular intervals, and convert them into stress data and deformation data according to the stress-strain relationship to form a segmented structure stress and deformation dataset that takes into account the time effect.

[0008] S13. Using a three-dimensional design model, construct a segmented structure simulation model corresponding to the feature segment under ideal conditions, set the load and boundary conditions under the idle state, and obtain the stress and deformation dataset of the segmented structure under ideal conditions through finite element analysis, including the stress and deformation data of each sampling point.

[0009] S14. Based on the stress and deformation dataset of the segmented structure considering the time effect obtained in S12, and the stress and deformation dataset of the segmented structure under ideal conditions obtained in S13, select an algorithm model for training, obtain the transmission relationship between the stress and deformation under ideal conditions and the stress and deformation of the segment after a certain time interval under this type of segmented structure, and obtain the inference model.

[0010] Optionally, the following steps may also be included:

[0011] S2. Determine the structural segments that need to be simulated and evaluated, and construct a segmented shelving finite element analysis model under ideal conditions;

[0012] S3. Using the finite element analysis model constructed in S2 as input to the simulation analysis and derivation model in S1, perform analysis and calculation considering the time effect based on the transfer relationship, obtain the segmented stress and deformation results under the action of time effect, and evaluate the structural safety of the segmented shelving state.

[0013] Optionally, in step S11, the ship type includes mainstream liquefied gas carriers, container ships, bulk carriers, and oil tankers, and typical sections include stern section, engine room section, bottom section, bilge section, side section, deck section, and bow section.

[0014] Optionally, in step S12, 10 sampling points are arranged in the length and width directions of the segment, and the strain data are measured every 12 hours.

[0015] Optionally, in step S14, the algorithm model includes neural networks, random forests, and kriging models.

[0016] Optionally, in step S14, a neural network algorithm is selected for training, specifically including:

[0017] (1) Determine the architecture, number of layers, number of neurons, and activation function of the neural network;

[0018] (2) The stress and deformation dataset of the segmented structure under ideal conditions obtained by the finite element analysis method of S13 is used as the input of the training dataset and submitted to the neural network;

[0019] (3) The stress and deformation data of the segmented structure considering the time effect collected by S12 are used as the output of the training dataset and submitted to the neural network;

[0020] (4) Based on the above definition of S13 and S12 data, construct the transfer relationship between the two types of data under ideal state and time effect, f=g(σFEA,Δt), and obtain the above derivation model; f represents the stress / deformation result considering the time effect, σFEA is the finite element stress / deformation result under ideal state; Δt is the shelving time.

[0021] Optionally, after step (4), the following steps are also included:

[0022] (5) Select several segments similar to the above feature segmentation structure as the test set for neural network training. Submit the stress and deformation data of these segments under ideal conditions as input to the deduction model formed in step (4) above to obtain the stress and deformation results under theoretical deduction time effect. Calculate the loss function between these results and the stress and deformation results under time effect obtained from the actual collection of the above segments to determine the difference between the two.

[0023] (6) Based on the loss function difference calculated in step (5), optimize the parameters of the deduction model. Through multiple calculations and parameter iteration optimization, make the loss function error within a reasonable range, and ensure that it can accurately describe the transmission law between stress and deformation data under ideal state and time effect. The deduction model is now complete.

[0024] As described above, this invention provides a simulation analysis method for ship hull section deformation under time-dependent conditions, aiming to solve the safety problem caused by the failure of traditional finite element analysis to account for the time factor in long-term storage scenarios, resulting in assessments that deviate from reality. This method acquires stress and deformation data of existing ship construction sections under different time periods of storage, and obtains ideal stress and deformation datasets of the section structure through a section structure simulation model. It then combines this with machine learning algorithms to construct a structural simulation analysis and deduction model under time-dependent conditions, thus incorporating the analysis of structural deformation under the time factor. This invention uses machine learning algorithms to obtain stress and deformation data of section sections under storage conditions considering the time effect. Building upon traditional finite element analysis, it further considers stress and deformation under the influence of time, providing a more comprehensive assessment of the safety of section storage, offering accurate judgment criteria for storage schemes, and ensuring the safety and reliability of the ship construction process. Attached Figure Description

[0025] Figure 1 The diagram shows a flowchart of the simulation analysis method in this invention.

[0026] Figure 2 The diagram shows the steps of step S1 in this invention. Detailed Implementation

[0027] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0028] It should be noted that the illustrations provided in this embodiment are only schematic representations of the basic concept of the present invention. Therefore, the illustrations only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0029] like Figure 1 As shown, this invention provides a simulation analysis method for the deformation of ship hull sections under suspension, considering the time effect, comprising the following steps:

[0030] S1. Construct structural simulation analysis and deduction models under time effects for typical structural sections of different ship types; such as... Figure 2 As shown, the specific steps include the following:

[0031] S11. Identify typical structural sections for different ship types, including mainstream liquefied gas carriers, container ships, bulk carriers, and oil tankers. Typical sections include stern section, engine room section, bottom section, bilge section, side section, deck section, and bow section.

[0032] S12. Select the existing segments as feature segments and construct a segmented structural stress and deformation database. Select key locations of the segments as sampling points, arrange strain sensors at the sampling points, and uniformly set strain acquisition time points within the segment rest time interval. Collect strain data from each strain sensor at regular intervals, and convert them into stress data and deformation data according to the stress-strain relationship to form a structured segmented structural stress and deformation dataset that considers the time effect.

[0033] For example, for a typical segmented structure of 10m×10m, 10×10 strain gauges are arranged in the length and width directions, for a total of 100 strain gauges. The strain data is measured every 12 hours. The measured strain data is converted into stress data and deformation data using the basic formulas of elasticity, thereby constructing a time-effect-based stress and deformation dataset for the segmented structure. This dataset can be recorded in tabular form, as shown in Table 1 below, and serves as the output for the subsequent construction of the structural simulation analysis and deduction model training dataset.

[0034] Table 1. Stress and Deformation Datasets for Piecewise Structures Considering Time Effects

[0035]

[0036] S13. Using the three-dimensional design model, construct the segmented structure simulation model corresponding to the feature segment under ideal conditions, set the load and boundary conditions under the idle state, and perform analysis to obtain the stress and deformation dataset of the segmented structure under ideal conditions, including the stress and deformation data of each sampling point.

[0037] The ideal segmented structure simulation model, which does not consider the influence of time effects, employs the traditional finite element method for structural analysis. This includes meshing the structural model with appropriately sized elements, adding material properties, setting boundary conditions for a static state, and applying gravity loads. The specific process can be as follows: discretize the segmented three-dimensional geometric model to determine reasonable element sizes and types; assign appropriate material properties to the structure, such as elastic modulus, Poisson's ratio, and density; set boundary constraints based on actual static conditions to simulate the contact relationship between the segments and supporting timbers or the foundation; and finally, apply gravity loads to reflect the structure's self-weight response in a static state. The stress and deformation results output by this model will serve as the baseline initial state for subsequent time-varying effect analysis.

[0038] Based on the above analysis, the stress and deformation values ​​at the strain gauge arrangement points (i.e., sampling points) in S12 are extracted to obtain the stress and deformation dataset of the segmented structure under ideal conditions without considering time. The dataset is as follows, and it serves as the input for the training dataset of the subsequent structural simulation analysis and deduction model:

[0039] Table 2. Dataset of stress and deformation of segmented structures under ideal conditions.

[0040]

[0041] S14. Based on the segmented structure stress and deformation dataset considering time effects obtained in S12, and the segmented structure stress and deformation dataset under ideal conditions obtained in S13, a suitable algorithm model is selected for training, including neural networks, random forests, and Kriging models. This yields the transmission relationship between the stress and deformation under ideal conditions and the stress and deformation of the segment after a certain time interval under this type of segmented structure, i.e., the deduction model described in S1. In other words, it learns the mapping relationship from the "ideal state" to the "actual state after a rest period," establishing the stress and deformation transmission law between the two.

[0042] The final inference model can be formally represented as:

[0043] f = g(σFEA, Δt)

[0044] Where f represents the stress / deformation result considering the time effect (after time-varying correction); σFEA is the finite element stress / deformation result under ideal conditions; Δt is the resting time.

[0045] Taking neural network algorithms as an example, the main steps are as follows:

[0046] (1) Determine the architecture, number of layers, number of neurons, and activation function of the neural network.

[0047] (2) The stress and deformation dataset of the segmented structure under ideal conditions obtained by the finite element analysis method of S13 is used as the input of the training dataset and submitted to the neural network;

[0048] (3) The stress and deformation data of the segmented structure considering the time effect collected by S12 are used as the output of the training dataset and submitted to the neural network;

[0049] (4) Based on the above definition of S13 and S12 data, construct the transmission relationship between the two types of data under ideal state and time effect, f=g(σFEA,Δt), that is, the deduction model described in S1.

[0050] For example, the stress at a certain measuring point under ideal conditions is 50 MPa, and it is left idle for 30 days; corresponding to the same measuring point, the measured stress value after 30 days is 55 MPa. The transitive relation is to obtain a function g such that for thousands of such input-output data pairs, g(50, 30) ≈ 55. As a simple example, f = a × σFEA + b × Δt;

[0051] (5) Select several segments similar to the above feature segmentation structure (their ideal state and time effect data are not in the above neural network training dataset) as the test set for neural network training. Submit the stress and deformation data of these segments under the ideal state as input to the deduction model formed in step (4) above to obtain the stress and deformation results under the theoretical deduction time effect. Calculate the loss function between these results and the stress and deformation results under the time effect actually collected by the above segments to determine the difference between the two.

[0052] (6) Based on the loss function difference calculated in step (5), the parameters of the simulation model are optimized. Through multiple calculations and parameter iteration optimization, the error of the loss function is kept within a reasonable range to ensure that the transmission law between stress and deformation data under ideal conditions and time effects can be accurately described. At this point, the structural simulation simulation model under time effects is completed. It should be noted that there are already many records of the algorithm principle and specific code in existing technologies, so they will not be repeated here.

[0053] S2. Determine the structural segments that need to be simulated and evaluated, and construct a segmented shelving finite element analysis model under ideal conditions;

[0054] S3. Using the finite element analysis model constructed in S2 as input to the simulation analysis and derivation model in S1, perform analysis and calculation considering the time effect based on the transfer relationship, obtain the segmented stress and deformation results under the action of time effect, and evaluate the structural safety of the segmented shelving state.

[0055] The above process will be described in detail below through examples.

[0056] Example 1

[0057] In this embodiment, a structural simulation evaluation considering the time effect is required for a certain type of deck section of the 99K VLEC under its resting state. The section size is 12×12m, which is similar to the section structure size used in the simulation model constructed in S1, and its resting method is also similar. The specific analysis process is as follows:

[0058] (1) Construct a finite element simulation model of the segment under ideal conditions, divide the segment into meshes of appropriate size, add material properties, set boundary conditions under the idle state, and apply gravity load. Then perform finite element calculations.

[0059] (2) Set 10×10 stress and deformation data extraction points in the length and width directions of the segment to obtain the stress and deformation dataset of the segmented structure under ideal conditions, as shown in Table 3 below:

[0060] Table 3 shows the stress and deformation data of the segmented structure under ideal conditions.

[0061]

[0062] (3) Using the above data as input, submit the structural simulation model of this type of ship section considering the time effect, and output the stress and deformation results considering the time effect, as shown in Table 4 below:

[0063] Table 4 shows the stress and deformation dataset of this segmented structure considering time effects.

[0064]

[0065] (4) Assuming that the segment needs to be shelved for 1 day according to the production plan, we extract 12 hours of data and compare the stress and deformation data with the allowable stress and allowable deformation. The allowable stress is determined according to the material properties of the segment steel plate. The segment is low carbon steel with an allowable stress of 235 MPa. The allowable deformation is taken as 1 / 800 of the segment length, i.e., 15 mm. If the analysis results do not exceed the allowable values, it means that the shelving state is safe. Otherwise, special reinforcement treatment is required.

[0066] In summary, this invention provides a simulation analysis method for the deformation of ship hull sections under temporary storage conditions, considering the time effect. This method aims to address the safety issue caused by the failure of traditional finite element analysis to account for the time factor in long-term storage scenarios, resulting in assessments that deviate from reality. The method acquires stress and deformation data of existing ship construction sections under different storage periods, and obtains ideal stress and deformation datasets of the sections through a structural simulation model. It then combines this with machine learning algorithms to construct a structural simulation analysis and deduction model under the time effect, thus incorporating the analysis of structural deformation under the time factor. This invention uses machine learning algorithms to obtain stress and deformation data of sections under temporary storage conditions considering the time effect. Building upon traditional finite element analysis, it further considers stress and deformation under the influence of time, providing a more comprehensive assessment of the safety of section storage, offering accurate judgment criteria for storage schemes, and ensuring the safety and reliability of the ship construction process.

[0067] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A simulation analysis method for the deformation of ship hull sections under suspension considering time effects, characterized in that, Includes the following steps: S1. Constructing structural simulation analysis and deduction models under time effects for typical structural sections of different ship types; specifically including: S11. Identify typical structural sections for different ship types; S12. Select the existing segments as feature segments, select key locations of the segments as sampling points, arrange strain sensors at the sampling points, uniformly set strain acquisition time points during the segment rest time interval, collect strain data of each strain sensor at regular intervals, and convert them into stress data and deformation data according to the stress-strain relationship to form a segmented structure stress and deformation dataset that takes into account the time effect. S13. Using a three-dimensional design model, construct a segmented structure simulation model corresponding to the feature segment under ideal conditions, set the load and boundary conditions under the idle state, and obtain the stress and deformation dataset of the segmented structure under ideal conditions through finite element analysis, including the stress and deformation data of each sampling point. S14. Based on the stress and deformation dataset of the segmented structure considering the time effect obtained in S12, and the stress and deformation dataset of the segmented structure under ideal conditions obtained in S13, select an algorithm model for training, obtain the transmission relationship between the stress and deformation under ideal conditions and the stress and deformation of the segment after a certain time interval under this type of segmented structure, and obtain the inference model.

2. The simulation analysis method for ship section deformation under suspension considering time effects according to claim 1, characterized in that, It also includes the following steps: S2. Determine the structural segments that need to be simulated and evaluated, and construct a segmented shelving finite element analysis model under ideal conditions; S3. Using the finite element analysis model constructed in S2 as input to the simulation analysis and derivation model in S1, perform analysis and calculation considering the time effect based on the transfer relationship, obtain the segmented stress and deformation results under the action of time effect, and evaluate the structural safety of the segmented shelving state.

3. The simulation analysis method for ship section deformation considering time effects according to claim 1, characterized in that: In step S11, the ship types include mainstream liquefied gas carriers, container ships, bulk carriers, and oil tankers, and typical sections include the stern section, engine room section, bottom section, bilge section, side section, deck section, and bow section.

4. The simulation analysis method for ship section deformation under suspension considering time effects according to claim 1, characterized in that: In step S12, 10 sampling points are arranged in the length and width directions of the segment, and the strain data are measured every 12 hours.

5. The simulation analysis method for ship section deformation under suspension considering time effects according to claim 1, characterized in that: In step S14, the algorithm models include neural networks, random forests, and kriging models.

6. The simulation analysis method for ship section deformation under suspension considering time effects according to claim 1, characterized in that: In step S14, a neural network algorithm is selected for training, specifically including: (1) Determine the architecture, number of layers, number of neurons, and activation function of the neural network; (2) The stress and deformation dataset of the segmented structure under ideal conditions obtained by the finite element analysis method of S13 is used as the input of the training dataset and submitted to the neural network; (3) The stress and deformation data of the segmented structure considering the time effect collected by S12 are used as the output of the training dataset and submitted to the neural network; (4) Based on the above definition of S13 and S12 data, construct the transfer relationship between the two types of data under ideal state and time effect, f=g(σFEA,Δt), and obtain the above derivation model; f represents the stress / deformation result considering the time effect, σFEA is the finite element stress / deformation result under ideal state; Δt is the shelving time.

7. The simulation analysis method for ship section deformation under suspension considering time effects according to claim 6, characterized in that: After step (4), the following is also included: (5) Select several segments similar to the above feature segmentation structure as the test set for neural network training. Submit the stress and deformation data of these segments under ideal conditions as input to the deduction model formed in step (4) above to obtain the stress and deformation results under theoretical deduction time effect. Calculate the loss function between these results and the stress and deformation results under time effect obtained from the actual collection of the above segments to determine the difference between the two. (6) Based on the loss function difference calculated in step (5), optimize the parameters of the deduction model. Through multiple calculations and parameter iteration optimization, make the loss function error within a reasonable range, and ensure that it can accurately describe the transmission law between stress and deformation data under ideal state and time effect. The deduction model is now complete.

Images