Intelligent ship type design method and system based on ship type value points

The ship design method employs a CVAE generative neural network to generate ship type value points from existing ship data, enhancing design efficiency and accuracy by leveraging performance attributes, verified through CFD simulation.

CN119089584AInactive Publication Date: 2024-12-06HUAZHONG UNIV OF SCI & TECH

Patent Information

Application Number
CN202411201316.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-29
Publication Date
2024-12-06
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The traditional ship design method based on the parent ship is inefficient due to its reliance on empirical support and lack of utilization of existing ship performance data, leading to prolonged optimization processes.

Method used

A ship design method and system utilizing a ship type value point generation model, specifically a conditional variational autoencoder (CVAE) generative neural network, that uses Froude number and ship residual resistance coefficient as inputs to generate ship type value points, which are then connected to form the ship's cross-section lines, leveraging existing ship data for efficient design.

Benefits of technology

The method achieves high-accuracy, time-efficient ship design by utilizing existing ship performance data, reducing reliance on prediction and optimization, and ensuring design accuracy through CFD simulation verification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089584A_ABST
    Figure CN119089584A_ABST
Patent Text Reader

Abstract

The invention belongs to the related technical field of ship design, and discloses a ship type intelligent design method and system based on ship type value points. The method comprises the following steps: off-line training: acquiring profile value points, Fourier numbers and ship residual resistance coefficients of a plurality of mother type ships as a training data set, and constructing a ship profile value point generation model taking the Fourier numbers and the ship residual resistance coefficients as input and the profile value points as output, training the type value point generation model by using the training data set; online application: inputting the Froude number of a ship to be designed and the residual resistance coefficient of the ship into the trained ship type value point generation model to obtain the type value point of the ship to be designed; and drawing the line type of the to-be-designed ship by using the obtained type value points of the to-be-designed ship. According to the invention, the technical problem that ship linetype design is irrelevant to ship performance data is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field related to ship design, and more specifically, relates to a ship type intelligent design method and system based on ship type value points. Background Art

[0002] Ship form design is a key link in ship design and an important manifestation of the innovation of ship design. At present, the modified mother prototype method is the main method of ship form design. This method draws on and inherits the excellent lines of the mother ship or series of ships. In the design process, the modified mother prototype method first determines the initial ship form based on the existing ship form data, and then repeatedly adjusts and calculates the initial ship form to make the ship performance meet the design requirements. In the modified mother prototype method, from the initial ship form to the final ship form, it is a process of continuously optimizing the ship form through the prediction and evaluation of ship performance. This process is not related to the existing ship data. In the ship form design process based on the modified mother prototype method, the mother ship only provides empirical support in determining the initial ship form. The optimization process from the initial ship form to the final ship form takes a lot of time and is the main factor leading to the low efficiency of ship form design.

[0003] The traditional design method based on the parent ship only uses the shape attribute data of the existing ship, and the performance attribute data of the existing ship is rarely used. If the shape and performance data of the parent ship can be comprehensively utilized to study a new design method that directly obtains the final ship form under given ship performance conditions, it can reduce the efficiency loss caused by the ship form optimization process in the traditional method and improve the efficiency of ship form design. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a method and system for intelligent ship design based on ship type value points, which solves the technical problem of low ship design efficiency caused by the fact that ship line design is irrelevant to ship performance data.

[0005] To achieve the above object, according to one aspect of the present invention, a ship type intelligent design method based on ship type value points is provided, the method comprising the following steps:

[0006] The Froude number and the ship residual resistance coefficient of the ship to be designed are input into the trained ship type value point generation model to obtain the ship type value point coordinates of the ship to be designed; the ship type value point generation model is obtained by the following method:

[0007] Collecting the coordinates of ship form value points, Froude numbers and ship residual resistance coefficients of multiple mother ships as training data sets, constructing a ship form value point generation model with the Froude number and the ship residual resistance coefficient as input and the form value point coordinates as output, and using the training data sets to train the ship form value point generation model;

[0008] Using the station corresponding to the horizontal coordinate of the ship type value point coordinates of the ship to be designed, the ship type value points corresponding to the same station are connected to form a surface. The surface is used as the cross-section line corresponding to the station. The cross-section lines corresponding to all stations are connected to form the line type of the ship to be designed, thereby realizing the design of the ship.

[0009] Further preferably, before the ship type value point coordinates of the mother ship are input into the ship type value point generation model, the ship type value point coordinates need to be normalized and dimensionless.

[0010] Further preferably, the coordinates of the ship type value point after normalization and dimensionless processing are expressed as follows:

[0011]

[0012] Among them, i is the station number, n is the ship type value point number under the station number, y i,n is the lateral coordinate of the ship type value point, z i,n is the vertical coordinate, L pp It is the length between perpendicular lines of the ship.

[0013] Further preferably, the ship type value point generation model adopts a conditional variational coding network generative neural network model.

[0014] Further preferably, the ship type value points corresponding to the same site are connected by spline curves.

[0015] According to another aspect of the present invention, a ship type intelligent design system based on ship type value points is provided, the system comprises a data acquisition module, a ship type value point generation model module, and a ship type generation module, wherein:

[0016] The data acquisition module is used to collect the coordinates of the ship type value points, the Froude number and the ship residual resistance coefficient of the mother ship;

[0017] The ship type value point generation model module is used to construct a ship type value point generation model, train the ship type value point generation model using a training data set, and obtain the ship type value points of the ship to be designed using the trained ship type value point generation model;

[0018] The ship type generation module is used to connect the ship type value points corresponding to the same site to form a curved surface, and to connect the cross-section lines corresponding to all sites to form the line type of the ship to be designed.

[0019] According to another aspect of the present invention, there is provided a ship type designed by using the above-mentioned ship type intelligent design method based on ship type value points.

[0020] According to another aspect of the present invention, there is provided a ship type intelligent design system based on ship type value points, the system comprising an actuator for executing the above-mentioned ship type intelligent design method based on ship type value points.

[0021] According to another aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned intelligent ship type design method based on ship type value points.

[0022] According to another aspect of the present invention, there is provided a computer program product, comprising a computer program / instruction, which implements the steps of the above-mentioned method when executed by a processor.

[0023] In general, the above technical solutions conceived by the present invention have the following beneficial effects compared with the prior art:

[0024] 1. The present invention constructs a type value point generation model and uses the model to obtain the type value points of the ship to be designed, connects the type value points under different station numbers with spline curves to obtain the line type of the ship to be designed, and the obtained line type is related to the ship parameters. The final obtained ship type has high accuracy, consumes time, and has high design efficiency;

[0025] 2. The present invention proposes a ship type intelligent design method based on ship type value points, selects conditional variational autoencoder as a generative neural network, and studies the ship type design method of CVAE generative neural network under given speed and ship residual resistance coefficient based on the parent ship data, and verifies the accuracy of the design results through CFD numerical simulation technology; Compared with the traditional ship type design method, the design method based on generative neural network makes full use of the resistance data of existing ships, reduces the dependence of ship type design on ship performance prediction and optimization technology, improves the efficiency of ship type design, and has important application value;

[0026] 3. In the present invention, the coordinates of the ship type value points are selected as the expression method of the ship type information, and the type value points of each station are standardized and dimensionless. These ship type value points can accurately describe the appearance characteristics of the ship, and are standardized and dimensionless. There is no need to consider the conversion of units during training and application. The Froude number and the ship residual resistance coefficient are used as input, and the type value points are used as output to generate a model from the ship type value points. The Froude number and the residual resistance coefficient can accurately reflect the resistance performance of the ship, and are also dimensionless quantities without any physical units. There is no need to consider the conversion of units during training and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a ship type intelligent design method based on ship type value points constructed according to a preferred embodiment of the present invention;

[0028] Figure 2 is a flow chart of the implementation process of the conditional variational autoencoder constructed according to the preferred embodiment of the present invention;

[0029] Figure 3 It is a schematic diagram of the ship type value point CVAE generative neural network structure constructed according to the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0031] A ship type intelligent design method based on ship type value points, the method comprising the following steps:

[0032] The coordinates of the ship's form points are usually a set of two-dimensional or three-dimensional coordinate points that describe the geometry of the hull. These points are usually located on the hull contour line, and when designing the hull, these points can be used to build mathematical models or perform ship simulation analysis. Usually, the coordinate values ​​of these points are arranged according to a certain pattern to facilitate the description of the shape characteristics and dimensions of the hull. The specific form and number of the coordinates of the ship's form points depend on the choice of the designer or design software. The form points are an intuitive representation of actual engineering, and are also an important basis for ship surface construction and ship resistance calculation, and contain a large amount of ship information.

[0033] Construct a sample library of the relationship between ship type value points and ship resistance performance. In this embodiment, a conditional variational auto encoder (CVAE) is used as a ship type value point generation model. CVAE can use the original data and the corresponding labels as the input of the encoder. The latent variables can capture the variation factors of one-to-many problems and can be used to generate data of a specified type, that is, the one-to-many property.

[0034] like Figure 2 As shown in the figure, the network structure of CVAE consists of two parts: encoder and generator (decoder). The input information is encoded by the encoder, the data is compressed in the latent space, and then restored by the decoder. Among them, the output data of the encoder is used as the input data of the decoder. CVAE adds data labels c based on the encoder input data and the decoder sampling data, so as to achieve the purpose of allowing the network model to generate data according to given attributes or categories.

[0035] Determine the coordinates of the ship type value points and resistance performance parameters. Figure 3 As shown in the figure, the ship type value points are selected as the input features of CVAE and expressed as follows:

[0036]

[0037] Froude number Fn and ship residual resistance coefficient C r As the input label of CVAE, the generative neural network structure of ship type value point CVAE is established. In the above parameters: i represents the station number, n represents the type value point number under the station number, y i,n Indicates the horizontal coordinate of the type value point, z i,n Indicates the vertical coordinate of the type value point, L pp Indicates the length between perpendicular lines of the ship.

[0038] The coordinates of the form value points of multiple mother ships are selected as sample data. The mother ship data is used to train the ship form value point CVAE generative neural network, and the ship form value points are generated under the conditions of residual resistance coefficient and Froude number of each test sample.

[0039] Since there is too much coordinate data information of the ship type value points, it is difficult to accurately infer the ship type value points based on the sample relationship. In order to better output the ship type value points based on the relationship sample library, three experimental methods are proposed:

[0040] 1) Increase the sample size. Some experiments have shown that the accuracy of the experimental results obtained after training the model with more experimental samples will be improved to a certain extent, but if the number of samples is too large, the time cost and computing resource consumption will be greater, and experimental comparison and selection are needed.

[0041] 2) Adjust the model structure. Adjust the model parameters or the decoding network to increase the output result enhancement structure. The preliminary experiment uses the feature enhancer to enhance the output results, better fit the original input features, and obtain more accurate coordinates of the ship type value points. This method has been confirmed in other fields.

[0042] 3) Considering that the model input is multidimensional data, it is planned to display the coordinates of the type value points in the form of three-dimensional data, crop the two-dimensional image with a certain position as the axis, connect the points on the same image in sequence, and form a two-dimensional image database of n samples. The problem of generating three-dimensional type value points is converted into multiple two-dimensional image problems, and then experimental verification is carried out.

[0043] The Froude number and the ship's residual resistance coefficient of the ship to be designed are input into the trained ship type value point generation model to obtain the coordinates of the type value points of the ship to be designed. The type value points at different station numbers are connected with spline curves to obtain the line type of the cross-section of the ship to be designed at the station number; repeating this process can obtain the line type of the cross-section of the ship to be designed at all station numbers, thereby realizing the design of the ship type.

[0044] like Figure 1 As shown in the figure, the coordinates of the ship type value points and the resistance information of the mother ship are used as sample data to train the ship type value point CVAE generative neural network, where the ship type value point coordinates are used as the input features of CVAE, the Froude number Fn and the ship residual resistance coefficient C r as the input label of CVAE; after the training, given the residual resistance coefficient and Froude number, the coordinates of the ship type value points can be obtained through the ship type value point CVAE generative neural network; the type value points at different station numbers are connected with spline curves to obtain the line type of the cross-section of the ship to be designed at the station number; repeating this process can obtain the line type of the cross-section of the ship to be designed at all station numbers, so as to realize the design of the ship type, and the three-dimensional shape surface of the designed ship can be obtained through the designed line type, and the CFD resistance numerical calculation of the three-dimensional model is performed. By comparing the CFD simulation results with the given residual resistance coefficient, the accuracy of the design results and the reliability of the ship type design method are further verified.

[0045] Then, based on CFD numerical simulation technology, the ship resistance under given Froude conditions is obtained. By comparing the CFD simulation results with the given residual resistance coefficient, the accuracy of the design results and the reliability of the ship design method are further verified.

[0046] It will be easily understood by those skilled in the art that the above description is only 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 in the protection scope of the present invention.

Claims

1. A ship type intelligent design method based on ship type value points, characterized in that: The method comprises the following steps: The Froude number and the ship residual resistance coefficient of the ship to be designed are input into the trained ship type value point generation model to obtain the ship type value point coordinates of the ship to be designed; the ship type value point generation model is obtained by the following method: Collecting the coordinates of ship form value points, Froude numbers and ship residual resistance coefficients of multiple mother ships as training data sets, constructing a ship form value point generation model with the Froude number and the ship residual resistance coefficient as input and the form value point coordinates as output, and using the training data sets to train the ship form value point generation model; Using the station corresponding to the horizontal coordinate of the coordinates of the ship type value point of the ship to be designed, the ship type value points corresponding to the same station are connected to form a surface. The surface is used as the cross-section line corresponding to the station. The cross-section lines corresponding to all stations are connected to form the line type of the ship to be designed, thereby realizing the design of the ship.

2. The ship type intelligent design method based on ship type value points according to claim 1, characterized in that: Before the ship type value point coordinates of the mother ship are input into the ship type value point generation model, the ship type value point coordinates are normalized and dimensionless.

3. A ship type intelligent design method based on ship type value points as claimed in claim 2, characterized in that: The coordinates of the ship type value point after normalization and dimensionless processing are expressed as follows: Among them, i is the station number, n is the ship type value point number under the station number, y i,n is the lateral coordinate of the ship type value point, z i,n is the vertical coordinate, L pp is the length between perpendicular lines of the ship, P i,n are the coordinates of the ship type value point.

4. A ship type intelligent design method based on ship type value points as claimed in claim 1 or 2, characterized in that: The ship type value point generation model adopts a conditional variational encoding network.

5. A ship type intelligent design method based on ship type value points as claimed in claim 1 or 2, characterized in that: The ship type value points corresponding to the same site are connected by spline curves.

6. A ship type intelligent design system based on ship type value points, characterized in that: The system includes a data acquisition module, a ship type value point generation model module, and a ship type generation module, among which: The data acquisition module is used to collect the coordinates of the ship type value points, the Froude number and the ship residual resistance coefficient of the mother ship; The ship type value point generation model module is used to construct a ship type value point generation model, train the ship type value point generation model using a training data set, and obtain the ship type value points of the ship to be designed using the trained ship type value point generation model; The ship type generation module is used to connect the ship type value points corresponding to the same site to form a curved surface, and to connect the cross-section lines corresponding to all sites to form the line type of the ship to be designed.

7. A ship type designed by using the ship type intelligent design method based on ship type value points as described in any one of claims 1 to 5.

8. A ship type intelligent design system based on ship type value points, characterized in that: The system includes an actuator, which is used to execute the ship type intelligent design method based on ship type value points as described in any one of claims 1-5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements a ship type intelligent design method based on ship type value points as described in any one of claims 1 to 5.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Citation Information

Patent Citations

  • Propeller model construction method based on ship test

    CN111914348A

  • Method for producing freeze-dried instant laver broth

    KR102285163B1

Cited By

  • Ship form design method and system

    CN120579273A