A knowledge-driven ship hull line intelligent design platform and method

Abstract

This invention belongs to, but is not limited to, the field of intelligent hull line design technology. It discloses a knowledge-driven intelligent hull line design platform and method. The user layer is used for users to input the main element requirements and key features of the designed ship. The application layer is used to retrieve similar initial hull types from a parent model library based on the user's design requirements, and then use a hull line knowledge inference model to obtain hull feature parameters with excellent resistance performance. The parametric geometric modeling software CAESES is then integrated to generate a three-dimensional geometric model of the hull surface using the parametric design module. The knowledge base layer is used to rationally and effectively acquire, represent, classify, and manage expert experience, actual ship design data, optimization simulation data, ship design principles, and regulatory knowledge, focusing on total resistance performance to establish a hull line design knowledge base. This invention can quickly generate inferred hull line schemes that meet design requirements and has significant engineering application value.

Description

Technical Field

[0001] This invention belongs to, but is not limited to, the field of intelligent hull design technology, and particularly relates to a knowledge-driven intelligent hull design platform and method. Background Technology

[0002] The hull lines determine a ship's sailing performance and are one of the most important aspects of overall ship design. With the development of computer technology, intelligent hull optimization methods based on Computational Fluid Dynamics (CFD) have become an important research direction in hull line design. Feng Baiwei et al., combining Simulation Based Design (SBD) technology, proposed a multi-objective hull line optimization platform. This platform integrates CFD numerical simulation technology, hull geometry reconstruction technology, optimization technology, and approximation technology, and used it to conduct hull line optimization research on various types of ships. Tran et al. used cubic spline interpolation functions to achieve Lackenby deformation of the hull shape and optimized the hull line of a fishing boat based on CFD, the Kriging surrogate model, and optimization techniques, achieving a reduction in total resistance of approximately 8.8%. Kim et al. selected 29 parameter variables to perform parametric geometric description of a small vessel and obtained the optimal hull shape by combining a deep neural network model and a genetic algorithm.

[0003] In recent years, the rapid development of artificial intelligence has opened up entirely new directions for the design field. Artificial intelligence requires a vast amount of knowledge for reasoning and decision-making, with knowledge engineering technology being a key focus. In the shipbuilding field, research on knowledge engineering technology currently mainly focuses on determining the ship's main dimensions, structural design, and layout design. Yang Shaoming applied knowledge engineering technology to solve the problem of multi-deck cabin layout, using production rule representation and object-oriented representation to represent the knowledge of ship living quarters layout, and completing knowledge reasoning according to cabin layout reasoning strategies. Cui et al., for the structural design of container ship cargo holds, used design rule methods and interpolation methods, drawing on past successful cases to guide the structural design of new ships. Zhao Tongming established a real-ship case library and a rule library containing design specifications, expert knowledge, etc., and realized the automatic and intelligent generation of parent ship selection and design schemes based on various reasoning algorithms. In the field of hull line design, some scholars have studied the implicit relationship between hull line parameters and ship performance, thus obtaining design knowledge of hull lines. Zhang Jianyi et al. constructed a knowledge graph of the hull lines of a straight-bow-twin-tail fin bulk carrier in the middle and lower reaches of the Yangtze River, which can reflect the variation law of different hull line parameters under different block coefficients and scale ratios. Ye Meng et al. used rough set theory to perform data mining on KCS container ship optimization simulation data, obtaining the implicit design rules between hull design variables and wave-making resistance. Zheng et al. extracted design knowledge from twin-fin hull optimization data through sensitivity analysis and self-organizing mapping neural networks, analyzing the implicit relationships between variables and between variables and the objective function.

[0004] Based on the above analysis, the urgent technical problems that need to be solved in the existing technology are: in the field of shipbuilding, the current research on knowledge engineering technology mainly focuses on the determination of ship main dimensions, structural design and layout design. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a knowledge-driven intelligent design platform and method for ship hull lines.

[0006] This invention is implemented as follows: a knowledge-driven intelligent design platform for ship hull lines, comprising a user layer, an application layer, and a knowledge base layer, with bidirectional data, information, and knowledge transfer between each layer;

[0007] The user layer is used for users to input the main design requirements of the ship, including length between perpendiculars L, beam B, draft T, and block coefficient C. B Fu Rude's number F r And the main characteristics of the designed ship, including ship type, bow and stern type, design speed, etc.;

[0008] The application layer is used to first search for similar initial ship types in the parent model library according to the user's design requirements, and then use the hull line knowledge reasoning model to further obtain the ship type characteristic parameters with excellent resistance performance; then integrate the parametric geometric modeling software CAESES to generate a three-dimensional geometric model of the hull surface using the parametric design module.

[0009] The knowledge base layer is used to acquire, represent, classify, and manage knowledge such as expert experience, actual ship design data, optimization simulation data, ship design principles, and standards in a reasonable and effective manner. It focuses on the overall resistance performance and establishes a knowledge base for hull line design.

[0010] Furthermore, the profile design knowledge base includes a profile knowledge graph, a parent model instance library, and a design rule library;

[0011] The knowledge graph of morphology is a curve showing the variation of several morphology parameters (longitudinal position of the center of buoyancy, inner fatness of the caudal fin, caudal shaft spacing ratio, and head width contraction ratio) with scale ratios L / B and B / T under different square coefficients.

[0012] The ship characteristic attributes used in the parent model instance library include ship number, ship name, ship type, bow shape, stern shape, length between perpendiculars, length-to-beam ratio, width-to-draft ratio, block coefficient, and design speed; the structural attributes corresponding to each instance consist of the parametric geometric model of that instance, the applicable hull line knowledge graph type, and the applicable design rule base type;

[0013] The profile design rule library includes cross section area curves, design waterline, head shape, and tail shape.

[0014] Furthermore, the important parameters for constructing the geometric model mainly include: (1) global geometric characteristic parameters: the length ratio of the parallel middle body, the length ratio of the inlet section, the length ratio of the outlet section, the longitudinal position of the center of buoyancy, etc.; (2) local geometric characteristic parameters: the inlet and outlet angles of the design waterline, the bulbous nose parameters, the tail fin parameters, etc.

[0015] Furthermore, the construction methods for the linear knowledge reasoning model include:

[0016] (1) Based on the input design requirements of the ship, retrieve similar ship type instances from the parent type instance library, determine the geometric parameters of the hull lines to be inferred based on the parametric model of the similar ship type, and realize the matching of the design ship with the hull line knowledge graph and the design rule library based on the applicable graph type and design rule library type of the similar ship type.

[0017] (2) Model reasoning and rule reasoning are used to obtain the global geometric feature parameter values ​​or value ranges of the hull lines in the corresponding hull line knowledge graph and design rule base respectively, so as to control the global hull line shape.

[0018] (3) Obtain the local geometric feature parameter values ​​or value range of the hull, control the local hull hull hull shape, and combine them with the global geometric feature parameters to form the hull ...

[0019] Furthermore, knowledge reasoning from the parent type instance library:

[0020] The Nearest Neighbor Search (NNS) algorithm, a commonly used method, is selected for retrieving mother ship instances; a weighted Euclidean distance formula is chosen, expressed as follows:

[0021]

[0022] Where: x i Let y be the i-th feature attribute of the design case; i Let be the i-th feature attribute of existing cases in the database; n is the number of types of feature attributes; w i Let be the weight of the i-th feature attribute.

[0023] Assuming each feature attribute has a 1 / n influence on the similarity score, the similarity between instances is calculated using the similarity formula:

[0024]

[0025] Furthermore, reasoning using the knowledge graph of shape lines:

[0026] By utilizing a gridded dataset that constructs a knowledge graph of the profile, and based on the variation patterns of profile parameters reflected in the knowledge graph, a step-by-step interpolation inference model is established for the profile parameters with respect to CB, L / B, and B / T. This allows for the inference of the graph parameter values ​​under actual CB, L / B, and B / T based on the graph curves. The interpolation formula is shown below:

[0027]

[0028] The inference model process is as follows:

[0029] 1) Select two adjacent Cs B (C B1 and C B2 ): For C B1 First, the actual B / T and L / B of the ship are determined based on the main scale elements provided by the user. Then, two adjacent L / B values ​​are selected in the hull lines knowledge graph, and two sets of parameter values ​​are obtained through interpolation: X skeg1* and X skeg2* ;

[0030] 2) Then, interpolate the actual L / B value to obtain the parameter value X corresponding to the square coefficient. skeg3 ;

[0031] 3) Similarly, for C B2 The parameter value X is obtained based on the actual L / B interpolation. skeg4 ;

[0032] 4) Ultimately based on X skeg3 and X skeg4 Interpolation yields the actual X skeg .

[0033] Furthermore, design a rule-based knowledge reasoning system:

[0034] The algorithm employs rule-based reasoning, matching problems based on production rules. The basic process is as follows: based on the design parameters input by the user, the rules stored in the rule base are evaluated one by one in sequence. If the rule is satisfied, the conclusion of that rule is output; otherwise, the rule is skipped until all rules have been evaluated.

[0035] Another objective of this invention is to provide a knowledge-driven intelligent hull line design method based on a knowledge-driven intelligent hull line design platform, comprising the following steps:

[0036] Step 1: The user inputs the main design requirements for the ship, including length between perpendiculars L, beam B, draft T, and block coefficient C. B Fu Rude's number F r And the main characteristics of the designed ship, including ship type, bow and stern type, design speed, etc.;

[0037] Step two: Based on the user's design requirements, firstly, similar initial ship types are searched in the parent model library. Then, the hull shape characteristic parameters with excellent resistance performance are obtained by using the hull line knowledge reasoning model. Next, the parametric geometric modeling software CAESES is integrated to generate a three-dimensional geometric model of the hull surface using the parametric design module.

[0038] Step 3 involves the reasonable and effective acquisition, representation, classification, and management of expert experience, actual ship design data, optimized simulation data, ship design principles, and standards, with a focus on total resistance performance, to establish a hull line design knowledge base.

[0039] Another object of the present invention is to provide a computer device, the computer device including a memory and a processor, the memory storing a computer program, and when the computer program is executed by the processor, causing the processor to perform the steps of the knowledge-driven intelligent hull line design method described above.

[0040] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0041] First, this invention proposes a knowledge-driven intelligent hull line design framework, mainly comprising three components: a user layer, an application layer, and a knowledge base. A knowledge base suitable for hull line design is constructed to store instance-based knowledge, hull line knowledge graphs, and rule-based knowledge. Multiple reasoning methods are employed for knowledge-based reasoning of hull line geometric parameters. Finally, a verification study is conducted using the hull line design of a 13,000-ton bulk carrier on the Yangtze River as an example. First, similar cases are retrieved from the parent model instance library to determine the applicable parametric model, hull line knowledge graph type, and design rule base type. Second, the global and local geometric feature parameters of the designed ship are inferred using the hull line knowledge graph model and design rule base, generating the inferred hull form. Finally, CFD verification shows that the inferred hull form exhibits good resistance performance, demonstrating that the knowledge-driven intelligent hull line design method can quickly generate inferred hull line schemes that meet design requirements, possessing significant engineering application value.

[0042] Secondly, by adopting this technology, the design cycle for ship hull lines can be shortened from the usual several weeks to just a few minutes. While ensuring the quality of hull line design, it significantly improves design efficiency and saves a considerable amount of time. This invention can be widely applied to various ship design and research and development units and has significant commercial value.

[0043] This invention differs from traditional hull design methods, which rely on expert experience or tank tests to determine the optimal solution. This process is time-consuming, and the resulting solution may not even be optimal. This invention proposes a knowledge-driven approach to the three-dimensional design of hull surfaces. By establishing a hull hull line knowledge base, it achieves knowledge-guided intelligent design, filling a technological gap in this field.

[0044] For years, people have hoped to use artificial intelligence to design ship hull lines, completely eliminating the tedious manual process. However, because hull surfaces cannot be mathematically described using geometric features, this field has not seen a significant breakthrough. This invention not only solves the problem of parametric expression of hull surfaces based on geometric feature parameters, but also breaks through the rapid design method of geometric feature parameters, truly realizing the automation and intelligence of hull line design, representing a revolutionary advancement in hull line design. Attached Figure Description

[0045] Figure 1 This is a knowledge-driven intelligent design framework diagram for ship hull lines provided in an embodiment of the present invention;

[0046] Figure 2 This is a structural diagram of the profile knowledge base provided in an embodiment of the present invention;

[0047] Figure 3 This is a parameter curve diagram of the profile knowledge graph provided in the embodiments of the present invention;

[0048] Figure 4 This is a framework diagram of the design rule base provided in an embodiment of the present invention;

[0049] Figure 5 This is a structural diagram of the inference model provided in an embodiment of the present invention;

[0050] Figure 6 This is a diagram of the knowledge graph reasoning model for the profile provided in this embodiment of the invention;

[0051] Figure 7 This is a flowchart of rule reasoning provided in an embodiment of the present invention;

[0052] Figure 8 This is an application flowchart of the intelligent profile design method provided in the embodiments of the present invention;

[0053] Figure 9 This is a geometric model of a 13,000-ton bulk carrier in the Yangtze River provided in this embodiment of the invention; (a) stern front view; (b) side view; (c) bottom bottom view;

[0054] Figure 10 This is a grid distribution diagram provided in an embodiment of the present invention;

[0055] Figure 11 This is provided by the embodiments of the present invention. Figure 11 Comparison of the lines of the proposed hull form (red dashed line) and the initial hull form (black solid line); (a) Comparison of transverse sections; (b) Comparison of longitudinal sections; (c) Comparison of waterlines;

[0056] Figure 12 These are comparative diagrams of hull surface pressure distribution provided in the embodiments of the present invention; (a) hull surface pressure distribution of the initial hull type; (b) hull surface pressure distribution of the inferred hull type. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to examples. It should be understood that the specific examples described herein are merely illustrative and not intended to limit the invention.

[0058] This invention combines knowledge engineering technology to establish a knowledge base and hybrid reasoning model applicable to the design of hull lines for cargo ships on the Yangtze River, and develops a knowledge-driven intelligent hull line design platform. Taking a 13,000-ton bulk carrier on the Yangtze River as an example, the knowledge-driven hull line design process is discussed, and finally, the hull line scheme obtained through reasoning is verified.

[0059] like Figure 1 The aforementioned knowledge-driven intelligent hull line design platform includes a user layer, an application layer, and a knowledge base layer, with bidirectional data, information, and knowledge transfer between each layer.

[0060] The user layer is used for users to input the main design requirements of the ship, including length between perpendiculars L, beam B, draft T, and block coefficient C. B Fu Rude's number F r And the main characteristics of the designed ship, including ship type, bow and stern type, design speed, etc.;

[0061] The application layer is used to first search for similar initial ship types in the parent model library according to the user's design requirements, and then use the hull line knowledge reasoning model to further obtain the ship type characteristic parameters with excellent resistance performance; then integrate the parametric geometric modeling software CAESES, and use the parametric design module to generate a three-dimensional geometric model of the hull surface.

[0062] The knowledge base layer is used to acquire, represent, classify, and manage knowledge such as expert experience, actual ship design data, optimization simulation data, ship design principles, and standards in a reasonable and effective manner. It focuses on the overall resistance performance and establishes a knowledge base for hull line design.

[0063] I. Construction of the Shape Line Knowledge Base

[0064] The knowledge base constructed in this invention includes a shape line knowledge graph, a parent model instance library, and a design rule library. This knowledge base is the foundation and core for subsequent knowledge reasoning. The knowledge base, as shown in the figure... Figure 2 As shown.

[0065] (1) This invention establishes an example library for Yangtze River cargo ship types, storing several parent ship design examples in the relational database MySQL. These design examples all adopt a parametric expression method and construct geometric models based on CAESES software. The important parameters for constructing the geometric model mainly include: (1) global geometric feature parameters: parallel midbody length ratio, inlet section length ratio, outlet section length ratio, longitudinal position of the center of buoyancy, etc.; (2) local geometric feature parameters: design waterline inlet and outlet angles, bulbous bow parameters, tail fin parameters, etc.

[0066] The ship characteristic attributes used in the parent model instance library include ship number, ship name, ship type, bowform, sternform, length between perpendiculars, length-to-beam ratio, beam-to-draft ratio, block coefficient, and design speed. The structural attributes corresponding to each instance consist of the instance's parametric geometric model, the applicable hull form knowledge graph type, and the applicable design rule base type. Subsequently, similar instances can be retrieved using SQL based on characteristic attributes to obtain the parametric geometric model and knowledge base types required for the design. The established parent model instance library is shown in Tables 1 and 2.

[0067] Table 1. Parent Type Instances (Characteristics)

[0068]

[0069] Table 2. Parent Type Instances (Structural Attributes)

[0070]

[0071] (2) Knowledge Graph of Shape Lines

[0072] The hull shape knowledge graph consists of curves showing the variation of several hull shape parameters (longitudinal position of the center of buoyancy, inner flank fatness of the caudal fin, tail shaft spacing ratio, and nose width contraction ratio) with scale ratios L / B and B / T under different square factors. These curves reflect the mapping relationship between the geometric characteristics of the hull shape and its overall drag performance. Some curves in the hull shape knowledge graph are shown below. Figure 3 As shown.

[0073] (3) Design rule base

[0074] This invention addresses the influence of hull geometry parameters on ship resistance and establishes a hull design rule library, such as... Figure 4 As shown.

[0075] Taking the description of the inlet angle at full load as an example, the knowledge representation using production rules is shown in Table 3. In the table, CP is the prismatic coefficient of the design ship, VS(kn) is the speed, LE(m) is the inlet length, and Fr is the Froude number.

[0076] Table 3 Recommended range of inlet angle for full-load waterline

[0077]

[0078] The range of values ​​for the inflow angle at full load can be represented by the following rule:

[0079] Rule 001: IF (CP > 0.78) THEN (30° < inlet angle < 40°)

[0080] Rule 002: IF (0.75 < CP < 0.78) AND (Fr < 0.182) THEN (26° < inlet angle < 28°)

[0081] ...

[0082] II. Linearity Knowledge Reasoning Model

[0083] This invention employs Case-Based Reasoning (CBR), Model-Based Reasoning (MBR), and Rule-Based Reasoning (RBR) for the parent model instance library, the model line knowledge graph, and the design rule base, respectively. The reasoning model established in this invention is as follows: Figure 5 As shown, it mainly consists of the following three steps:

[0084] 1) Based on the input design requirements of the ship, retrieve similar ship type instances from the parent model instance library, determine the hull geometry parameters to be inferred based on the parametric model of the similar ship type, and achieve matching between the design ship and the hull knowledge graph and design rule library based on the applicable graph type and design rule library type of the similar ship type.

[0085] 2) Model reasoning and rule reasoning are used to obtain the global geometric feature parameter values ​​or value ranges of the hull lines from the corresponding hull line knowledge graph and design rule base, respectively, to control the global hull line shape.

[0086] 3) Obtain the local geometric feature parameter values ​​or value ranges of the hull lines, control the local hull line shape, and combine them with the global geometric feature parameters to form the hull line parameter scheme of the designed ship.

[0087] (1) Knowledge reasoning of the parent instance library

[0088] This invention selects the commonly used Nearest Neighbor Search (NNS) method for retrieving mother ship instances. The core of the NNS method lies in calculating concept distance, assigning weights, and calculating similarity. This invention chooses a weighted Euclidean distance formula, expressed as:

[0089]

[0090] Where: x i Let y be the i-th feature attribute of the design case; i Let be the i-th feature attribute of existing cases in the database; n is the number of types of feature attributes; w i Let be the weight of the i-th feature attribute.

[0091] Assuming each feature attribute has a 1 / n influence on the similarity score, the similarity between instances is calculated using the similarity formula:

[0092]

[0093] (2) Reasoning based on the knowledge graph of the shape line:

[0094] By utilizing a gridded dataset that constructs a knowledge graph of the profile, and based on the variation patterns of profile parameters reflected in the knowledge graph, a step-by-step interpolation inference model is established for the profile parameters with respect to CB, L / B, and B / T. This allows for the inference of the graph parameter values ​​under actual CB, L / B, and B / T based on the graph curves. The interpolation formula is shown below:

[0095]

[0096] The atlas parameter is the hypertrophy of the inner side of the caudal fin X. skegFor example, the reasoning model is as follows Figure 6 As shown, the main process is as follows:

[0097] 1) Select two adjacent Cs B (C B1 and C B2 ): For C B1 First, the actual B / T and L / B of the ship are determined based on the main scale elements provided by the user. Then, two adjacent L / B values ​​are selected in the hull lines knowledge graph, and two sets of parameter values ​​are obtained through interpolation: X skeg1* and X skeg2* ;

[0098] 2) Then, interpolate the actual L / B value to obtain the parameter value Xskeg3 corresponding to the square coefficient;

[0099] 3) Similarly, for C B2 The parameter value Xskeg4 is obtained based on the actual L / B interpolation.

[0100] 4) Ultimately based on X skeg3 and X skeg4 Interpolation yields the actual X skeg .

[0101] (3) Design rule base knowledge reasoning

[0102] The rule base reasoning design employs rule-based reasoning, matching problems based on production rules. The basic process is as follows: based on the design parameters input by the user, the rules stored in the rule base are evaluated sequentially. If a rule is satisfied, its conclusion is output; otherwise, the rule is skipped, until all rules have been evaluated. The flowchart is shown below. Figure 7 As shown.

[0103] The specific reasoning process is as follows: the user inputs initial conditions such as the main dimensions, block coefficient, and Froude number of the designed ship, and the computer sequentially matches these conditions against the design rules for geometric parameters in the library. Taking the design rule for the full-load waterline inlet angle as an example, if the design ship's C... P 0.76, F r The value is 0.18. First, match rule 001. If the condition (C) is not met, the match is terminated. P >0.78), discard; match rule 002, if its condition (0.75 < C) is met. P <0.78 and F r If the angle is less than 0.182, then the output rule 002 conclusion (26° < inlet angle < 28°) is obtained, thus completing the reasoning for the geometry parameters of this profile. Figure 4 After performing similar matching on all the profile geometry parameters in the design rule base shown, the final inference result of the design rule base can be obtained.

[0104] Evidence related to the technical effects obtained from the examples of this invention.

[0105] This invention uses the hull line design of a 13,000-ton bulk carrier on the Yangtze River as an example to illustrate the feasibility of a knowledge-driven intelligent hull line design method. The specific process is as follows: Figure 8 As shown.

[0106] 1. Generation of hull lines

[0107] The Yangtze River 13,000-ton bulk carrier is a straight-bow, double-stern vessel with L / B = 5.8257 and B / T = 3.9636. The main parameters of the vessel are shown in Table 4.

[0108] Table 4. Main parameters of the 13,000-ton bulk carrier on the Yangtze River

[0109]

[0110] The reasoning process is as follows:

[0111] 1) Based on the listed ship type characteristics and main elements, first perform instance reasoning to derive the applicable parametric model, hull line knowledge graph type, and design rule base type. The similarity of similar ship types in the instance library is shown in Table 5. Select the ship type with the highest similarity, and recommend the straight bow-twin fin cargo ship hull line knowledge graph. Select the cargo ship rule base for the design rule base.

[0112] Table 5. Similarity of Example Ship Types

[0113]

[0114] 2) Based on the knowledge graph of straight-bow-twin-fin cargo ship hull lines and the cargo ship design rule base, the input data are the main dimensions, block coefficient and Froude number of the designed ship, and the global geometric feature parameter values ​​or value ranges of the hull lines are obtained, as shown in Table 6.

[0115] Table 6 Global Geometric Feature Parameters

[0116]

[0117]

[0118] 3) Obtain the local geometric feature parameter values ​​or value ranges of the profile, as shown in Table 7.

[0119] Table 7 Local geometric feature parameters

[0120]

[0121] 4) Considering the arrangement constraints proposed by the shipowner (such as stern shaft height, gearbox arrangement, etc.) and the deformation coupling of some parameters, some suitable global and local geometric feature parameters are selected from Tables 6 and 7. The parameter values ​​are the median values ​​of the range. Combined with the principal dimension parameters, a parameterized hull line scheme is generated, as shown in Table 8.

[0122] Table 8 Parametric Ship Type Scheme

[0123]

[0124]

[0125] 2. Geometric model generation and drag performance calculation

[0126] Substituting the parameterized hull lines scheme into the parameterized geometric model derived from the example reasoning, the final generated geometric model is as follows: Figure 9 As shown.

[0127] The geometric models of the initial and inferred hull forms were scaled at a ratio of 1:25, and the total resistance at the model scale was calculated using the fully viscous flow CFD software STAR-CCM+. The computational domain's forward boundary was set to 1.5 times the ship's length, and the aft boundary to 2.5 times the ship's length. Mesh refinement was applied to the hull surface, surrounding area, free surface, and Kelvin wave locations, while ensuring the Y+ value on the hull surface remained within a normal range. The final generated mesh size was approximately 1.8 million, as shown in the mesh distribution diagram below. Figure 10 As shown. The k-ε model was used for the turbulence model, with a time step of 0.02 s. The total resistance was numerically simulated for the initial and inference hull forms at a design speed of 18.5 km / h.

[0128] 3. Comparative Analysis of Resistance Performance of Inference Ship Types

[0129] Table 9 shows a comparison of the resistance calculation results and some hydrostatic data for the initial and inference hull forms of the Yangtze River 13,000-ton bulk carrier. The total resistance of the inference hull form model is 3.06% lower than that of the initial hull form model. Figure 11 This section compares the hull lines of the initial and proposed ship designs. Figure 12 A comparison of hull surface pressure distribution diagrams is provided. The data in the table shows that the displacement volume of the indirect hull type is slightly larger, but the wetted surface area is reduced, resulting in a slight decrease in frictional resistance. However, the decrease is not significant; therefore, the decrease in total resistance is mainly due to changes in residual resistance. Analysis Figure 12 It can be seen that, for the pressure distribution at the bow, the high-pressure area of ​​the indirect-flow hull type is smaller, while the absolute value of the low-pressure area at the bottom is not large. Therefore, the pressure difference at the bow of the indirect-flow hull type is smaller, and the pressure distribution is more uniform. Similarly, for the pressure distribution at the stern, the pressure difference of the indirect-flow hull type is also smaller, and the pressure distribution is more uniform.

[0130] By combining the inferred hull geometry parameters and comparing the hull lines of the inferred and initial hull forms, it can be seen that there are significant differences in the bow and stern lines of the two hull forms. Due to knowledge-based inference of bow parameters such as the minimum beam-to-height ratio of the stealthy bulbous bow (inferred value 0.8), the transverse section area coefficient at the bow rise (inferred value 0.345), the curvature of the stealthy bulbous bow (inferred value 0.91), and the bow width contraction ratio (inferred value 0.9065), the concave portion of the stealthy bulbous bow in the deformed inferred hull form becomes more pronounced, the bow is more slender, and the U-shape of the bow transverse section is greater, which can reduce wave-making resistance. Similarly, for the stern, after reasoning about the tunnel inclination angle (inferred value 18), stern cap height ratio (inferred value 0.7), stern shaft spacing ratio (inferred value 0.5225), inner stern fin bulkiness (inferred value 0.5709), and outer stern fin bulkiness (inferred value 0.7), the inferred hull form's stern cross section is closer to a V-shape, which is beneficial for water flow along the longitudinal direction and less prone to separation, thus improving drag performance. However, the initial hull form did not specifically design these hull parts and hull geometry parameters (based on drag performance), leaving considerable room for hull line design optimization.

[0131] Table 9 Comparison of hydrostatic and resistance results

[0132]

[0133] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A knowledge-driven intelligent design platform for ship hull lines, characterized in that, It includes a user layer, an application layer, and a knowledge base layer, and each layer can exchange data, information, and knowledge bidirectionally. The user layer is used for users to input the main design requirements of the ship, including length between perpendiculars L, beam B, draft T, and block coefficient C. B Fu Rude's number F r And the main characteristics of the designed ship, including ship type, bow and stern type, and design speed; The application layer is used to first search for similar initial ship types in the parent model library according to the user's design requirements, and then use the line knowledge reasoning model to obtain the key parameters of the ship type that meet the resistance performance; then integrate the parametric geometric modeling software CAESES to generate a three-dimensional geometric model of the hull surface using the parametric design module. The knowledge base layer is used to acquire, represent, classify and manage expert experience, actual ship design data, optimization simulation data, ship design principles and standard knowledge in a reasonable and effective manner. With total resistance performance as the key focus, a knowledge base for hull line design is established. The profile design knowledge base includes a profile knowledge graph, a parent model instance library, and a design rule library; The shape knowledge graph consists of curves showing the variation of several shape parameters with scale ratios L / B and B / T under different square coefficients. The shape parameters include the longitudinal position of the center of buoyancy, the inner fatness of the caudal fin, the caudal shaft spacing ratio, and the head width contraction ratio. The ship characteristic attributes used in the parent model instance library include ship number, ship name, ship type, bow shape, stern shape, length between perpendiculars, length-to-beam ratio, width-to-draft ratio, block coefficient, and design speed; the structural attributes corresponding to each instance consist of the parametric geometric model of that instance, the applicable hull line knowledge graph type, and the applicable design rule base type; The profile design rule library includes cross section area curves, design waterline types, head shape, and tail shape; The methods for constructing a knowledge reasoning model for shape lines include: (1) Based on the input design requirements of the ship, retrieve similar ship type instances from the parent type instance library, determine the geometric parameters of the hull lines to be inferred based on the parametric model of the similar ship type, and realize the matching of the design ship with the hull line knowledge graph and the design rule library based on the applicable graph type and design rule library type of the similar ship type. (2) Model reasoning and rule reasoning are used to obtain the global geometric feature parameter values ​​or value ranges of the hull lines from the corresponding hull line knowledge graph and design rule base respectively, so as to control the global hull line shape. (3) Obtain the local geometric feature parameter values ​​or value range of the hull, control the local hull hull hull shape, and combine them with the global geometric feature parameters to form the hull ... Mother type instance library knowledge reasoning: The commonly used nearest neighbor search method is selected for retrieving mother ship instances; a weighted Euclidean distance formula is chosen, expressed as follows: ； Where: x i Let y be the i-th feature attribute of the design case; i Let be the i-th feature attribute of existing cases in the database; n is the number of types of feature attributes; w i Let be the weight of the i-th feature attribute; Assuming each feature attribute has a 1 / n influence on the similarity score, the similarity between instances is calculated using the similarity formula: ； Linearity Knowledge Graph Reasoning: By utilizing a gridded dataset that constructs a knowledge graph of the profile, and based on the variation patterns of profile parameters reflected in the knowledge graph, a step-by-step interpolation inference model is established for the profile parameters with respect to CB, L / B, and B / T. This allows for the inference of the graph parameter values ​​under actual CB, L / B, and B / T based on the graph curves. The interpolation formula is shown below: ； The inference model process is as follows: 1) Select two adjacent Cs B (C B1 and C B2 ): For C B1 First, the actual B / T and L / B of the ship are determined based on the main scale elements provided by the user. Then, two adjacent L / B values ​​are selected in the hull lines knowledge graph, and two sets of parameter values ​​are obtained through interpolation: X skeg1* and X skeg2* ; 2) Then, interpolate the actual L / B value to obtain the parameter value X corresponding to the square coefficient. skeg3 ; 3) Similarly, for C B2 The parameter value X is obtained based on the actual L / B interpolation. skeg4 ; 4) Ultimately based on X skeg3 and X skeg4 Interpolation yields the actual X skeg ; Design a rule base for knowledge reasoning: The algorithm employs rule-based reasoning, matching problems based on production rules. The basic process is as follows: based on the design parameters input by the user, the rules stored in the rule base are evaluated one by one in sequence. If the rule is satisfied, the conclusion of that rule is output; otherwise, the rule is skipped until all rules have been evaluated.

2. The knowledge-driven intelligent hull line design platform as described in claim 1, characterized in that, Important parameters for constructing a geometric model include: (1) Global geometric feature parameters: length ratio of parallel midbody, length ratio of inlet section, length ratio of outlet section, and longitudinal position of the center of buoyancy; (2) Local geometric feature parameters: inlet and outlet angles of the design waterline, bulb nose parameters, and tail fin parameters.

3. A knowledge-driven intelligent hull line design method based on the knowledge-driven intelligent hull line design platform as described in any one of claims 1 to 2, characterized in that, Includes the following steps: Step 1: The user inputs the main requirements of the design vessel, including length between perpendiculars L, beam B, draft T, block coefficient CB, Froude number Fr, and the main characteristics of the design vessel, including vessel type, bow and stern type, and design speed. Step two: Based on the user's design requirements, firstly, similar initial ship types are searched in the parent model library. Then, the key parameters of the ship type that meet the resistance performance are obtained by using the hull line knowledge reasoning model. Then, the parametric geometric modeling software CAESES is integrated to generate a three-dimensional geometric model of the hull surface using the parametric design module. Step 3 involves the reasonable and effective acquisition, representation, classification, and management of expert experience, actual ship design data, optimized simulation data, ship design principles, and standard knowledge, with a focus on total resistance performance, to establish a hull line design knowledge base.

4. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, it causes the processor to perform the steps of the knowledge-driven intelligent hull line design method as described in any one of claims 1 to 2.

5. A computer-readable storage medium, characterized in that, The device stores a computer program, which, when executed by a processor, causes the processor to perform the steps of the knowledge-driven intelligent hull line design method as described in any one of claims 1 to 2.

Images