Ship Optimization Design Method, Device and Storage Medium

Abstract

The present invention relates to the fields of equipment manufacturing and computer technology, and specifically discloses a ship optimization design method, device, and storage medium, including: determining ship hydrodynamic performance target information according to ship design objectives; inputting the ship hydrodynamic performance target information into a pre-trained ship hull form optimization model to obtain ship hull form target parameters matching the ship hydrodynamic performance target information; and completing ship design according to the ship hull form target parameters; wherein the pre-trained ship hull form optimization model is obtained by training a neural network model based on multiple groups of corresponding ship form parameters and hydrodynamic calculation results and optimizing it through a heuristic optimization algorithm. The ship optimization design method provided by the present invention improves the hull form design efficiency and thus enhances the ship design efficiency.

Description

Technical Field

[0001] The present invention relates to the field of equipment manufacturing and computer technology, and in particular to a ship optimization design method, a ship optimization design device and a storage medium. Background Art

[0002] In the early stages of ship design, designers usually transform the lines of a standard mother ship by stretching, compressing and other operations on the mother ship's lines, and then calculate the wave-making resistance coefficient of the new lines. The entire design process is cumbersome and repetitive, and consumes a lot of time.

[0003] The new line needs to calculate the hydrodynamic performance. When using commercial software for hydrodynamic calculation, it takes a long time, and repeated settings are required under different working conditions. The calculation progress cannot be monitored and the entire calculation process is uncontrollable. When using intelligent optimization algorithms for optimization calculations, it involves population iteration, and the time-consuming simulation calculation will be highlighted. How to realize the automatic operation of the calculation and shorten the hydrodynamic calculation time is particularly important.

[0004] Therefore, how to improve the efficiency of mid-line design in ship design has become a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0005] The present invention provides a ship optimization design method, a ship optimization design device and a storage medium, which solve the problem of low ship design efficiency caused by low mid-line design efficiency in ship design existing in the related art.

[0006] As a first aspect of the present invention, a ship optimization design method is provided, which comprises:

[0007] Determine the target information of ship hydrodynamic performance according to the ship design objectives;

[0008] Inputting the ship hydrodynamic performance target information into a pre-trained ship line optimization model to obtain ship line target parameters matching the ship hydrodynamic performance target information;

[0009] Complete ship design according to the target ship line parameters;

[0010] The pre-trained ship line optimization model is obtained by training a neural network model based on multiple groups of corresponding ship type parameters and hydrodynamic calculation results and optimizing through a heuristic optimization algorithm.

[0011] Furthermore, the pre-trained ship line optimization model is obtained by training a neural network model and optimizing it through a heuristic optimization algorithm according to multiple groups of corresponding ship type parameters and hydrodynamic calculation results, including:

[0012] Obtain the design matrix according to the variable control points of the ship;

[0013] Obtain the deformed new hull lines after deforming the hull lines of the ship's mother ship through radial basis functions;

[0014] Perform iterative calculations based on the design matrix and the hull parameters of the deformed new hull lines to obtain the hydrodynamic calculation results;

[0015] Train through the neural network model according to the hull parameters of the deformed new hull lines and the hydrodynamic calculation results to obtain the neural network model to be optimized;

[0016] Perform optimization calculations on the neural network model to be optimized according to the heuristic optimization algorithm to obtain the pre-trained ship hull line optimization model.

[0017] Further, obtaining the design matrix according to the variable control points of the ship includes:

[0018] Determine at least some of the moving control points in the bow moving control points as the variable control points of the ship, where the ship is divided into ship fixed control points and bow moving control points;

[0019] Determine the constraint range of the variable control points of the ship;

[0020] Perform stratified sampling through Latin hypercube sampling according to the constraint range of the variable control points of the ship to obtain the design matrix.

[0021] Further, the hull parameters of the deformed new hull lines obtained after deforming the hull lines of the ship's mother ship through radial basis functions include:

[0022] Parametrically express the hull lines of the ship's mother ship, and parametrically express both the ship fixed control points and the bow moving control points;

[0023] Perform variation processing on the design variables that are variable control points in the bow moving control points according to the radial basis function to obtain the deformed new hull lines.

[0024] Further, performing iterative calculations based on the design matrix and the hull parameters of the deformed new hull lines to obtain the hydrodynamic calculation results includes:

[0025] Determine the hull parameters of the deformed new hull lines and the number of iterative steps, where the hull parameters include the ship length, the molded depth, and the molded breadth;

[0026] Perform iterative calculations on the design matrix according to the preset number of iterative steps to obtain the hydrodynamic calculation results, and the hydrodynamic calculation results at least include the wave-making resistance of the hull and the fluid characteristics around the hull.

[0027] Further, the ship type parameters of the deformed new line and the hydrodynamic calculation results are used to train a neural network model to obtain a neural network model to be optimized, including:

[0028] Determine the optimization target in the hydrodynamic calculation results, where the optimization target in the hydrodynamic calculation results at least includes the wave-making resistance of the hull and / or the fluid characteristics around the hull;

[0029] Train the neural network model according to the optimization target in the hydrodynamic calculation results and the ship type parameters of the deformed new line to obtain a neural network model to be optimized.

[0030] Further, train the neural network model according to the optimization target in the hydrodynamic calculation results and the ship type parameters of the deformed new line to obtain a neural network model to be optimized, including:

[0031] Screen the optimization target in the hydrodynamic calculation results according to a preset screening condition to obtain valid data of the optimization target, where the preset screening condition at least includes prior knowledge and design specifications;

[0032] Train the neural network model according to the valid data of the optimization target and the ship type parameters of the deformed new line to obtain a neural network model to be optimized.

[0033] Further, perform an optimization calculation on the neural network model to be optimized according to the heuristic optimization algorithm to obtain a pre-trained ship hull line optimization model, including:

[0034] Determine design variables according to the variable control points of the ship, and determine target variables according to the optimization target in the hydrodynamic calculation results;

[0035] Determine the algorithm type of the heuristic optimization algorithm and the algorithm parameters corresponding to the algorithm type, where the algorithm type includes any one of genetic algorithm, simulated annealing method, and ant colony algorithm, and the algorithm parameters include population, crossover rate, and mutation rate;

[0036] Call the neural network model to be optimized according to the determined algorithm type, algorithm parameters, design variables, and target variables to perform an optimization calculation to obtain a pre-trained ship hull line optimization model.

[0037] As another aspect of the present invention, there is provided a ship optimization design device for implementing the ship optimization design method described above, where it includes:

[0038] A determination module for determining ship hydrodynamic performance target information according to the ship design target;

[0039] An acquisition module, configured to input the ship hydrodynamic performance target information into a pre-trained ship hull form optimization model, and obtain ship hull form target parameters matching the ship hydrodynamic performance target information;

[0040] A design module, configured to complete ship design according to the ship hull form target parameters;

[0041] Wherein, the pre-trained ship hull form optimization model is obtained by training a neural network model according to multiple groups of corresponding ship form parameters and hydrodynamic calculation results and optimizing it through a heuristic optimization algorithm.

[0042] As another aspect of the present invention, a storage medium is provided, which is used to store computer instructions, and when the computer instructions are loaded and executed by a processor, the ship optimization design method described above is implemented.

[0043] The ship optimization design method provided by the present invention determines ship hydrodynamic performance target information according to the ship design target, and then obtains ship hull form target parameters matching the ship hydrodynamic performance target information based on the pre-trained ship hull form optimization model. Finally, ship design can be completed based on the ship hull form target parameters. Before this ship design, a neural network model is trained with multiple groups of corresponding ship form parameters and hydrodynamic calculation results, and then a pre-trained ship hull form optimization model is obtained through a heuristic optimization algorithm. The obtained pre-trained ship hull form optimization model can be directly used to obtain ship hull form target parameters. This ship optimization design method can effectively improve the ship design efficiency, and since the pre-trained ship hull form optimization model is trained based on multiple groups of matching ship form parameters and hydrodynamic calculation results and optimized through a heuristic optimization algorithm, it can also improve the reliability of ship design. Description of the Drawings

[0044] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention, but do not constitute a limitation to the present invention.

[0045] Figure 1 It is a flowchart of the ship optimization design method provided by the present invention.

[0046] Figure 2 It is a flowchart for obtaining the pre-trained ship hull form optimization model provided by the present invention.

[0047] Figure 3a It is a schematic diagram of ship fixed point division provided by the present invention.

[0048] Figure 3b It is a flowchart for obtaining the design matrix provided by the present invention.

[0049] Figure 4 Schematic diagram of Latin hypercube sampling provided by the present invention.

[0050] Figure 5 Flow chart of Latin hypercube sampling provided by the present invention.

[0051] Figure 6 Schematic diagram of classification of basis function expressions provided by the present invention.

[0052] Figure 7 Flow chart of obtaining the hull parameters of the new deformed line provided by the present invention.

[0053] Figure 8 Flow chart of obtaining the hydrodynamic calculation results provided by the present invention.

[0054] Figure 9 Hydrodynamic calculation flow chart provided by the present invention.

[0055] Figure 10 Flow chart of obtaining the neural network model to be optimized provided by the present invention.

[0056] Figure 11 Schematic diagram of the three-layer perception model of the neural network model provided by the present invention.

[0057] Figure 12 Flow chart of the heuristic optimization algorithm provided by the present invention.

[0058] Figure 13 Flow chart of using the heuristic optimization algorithm to optimize the neural network model to be optimized to obtain the optimal hull line provided by the present invention.

[0059] Figure 14 Structure block diagram of the ship optimization design device provided by the present invention.

[0060] Figure 15 Structure block diagram of the electronic device provided by the present invention. Detailed implementation manners

[0061] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0062] To enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0063] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned accompanying drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so as to describe the embodiments of the present invention here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0064] In this embodiment, a ship optimization design method is provided. Figure 1 It is a flowchart of the ship optimization design method provided according to the embodiment of the present invention, as Figure 1 shown, including:

[0065] S100. Determine the ship hydrodynamic performance target information according to the ship design target;

[0066] In the embodiment of the present invention, the ship design target can be specifically understood as the performance requirements that the ship needs to achieve, mainly including hydrodynamic performance. Therefore, based on this ship design target, the ship hydrodynamic performance target information can be determined, that is, the ship hydrodynamic performance target information required for the current ship design can be determined.

[0067] S200. Input the ship hydrodynamic performance target information into a pre-trained ship hull form optimization model to obtain ship hull form target parameters matching the ship hydrodynamic performance target information;

[0068] After determining the ship hydrodynamic performance target information, input this ship hydrodynamic performance target information as input information into the pre-trained ship hull form optimization model. Since this pre-trained ship hull form optimization model is trained based on multiple groups of corresponding ship form parameters and hydrodynamic calculation results, when the ship hydrodynamic performance target information is input, ship hull form target parameters corresponding to this ship hydrodynamic performance target information can be obtained. Based on this ship hull form target parameter, ship design can be realized.

[0069] S300. Complete the ship design according to the target parameters of the ship hull lines;

[0070] Among them, the pre-trained ship hull line optimization model is obtained by training a neural network model based on multiple groups of corresponding ship type parameters and hydrodynamic calculation results and optimizing it through a heuristic optimization algorithm.

[0071] The ship optimization design method provided by the present invention determines the target information of the ship's hydrodynamic performance according to the ship design target, and then obtains the target parameters of the ship hull lines matching the target information of the ship's hydrodynamic performance based on the pre-trained ship hull line optimization model. Finally, the ship design can be completed based on the target parameters of the ship hull lines. Before this ship design, a neural network model is trained with multiple groups of corresponding ship type parameters and hydrodynamic calculation results, and then a pre-trained ship hull line optimization model is obtained through a heuristic optimization algorithm. The obtained pre-trained ship hull line optimization model can be directly used to obtain the target parameters of the ship hull lines. This ship optimization design method can effectively improve the ship design efficiency, and because the pre-trained ship hull line optimization model is trained based on multiple groups of matching ship type parameters and hydrodynamic calculation results and optimized through a heuristic optimization algorithm, it can also improve the reliability of the ship design.

[0072] In the embodiment of the present invention, the pre-trained ship hull line optimization model is obtained by training a neural network model based on multiple groups of corresponding ship type parameters and hydrodynamic calculation results and optimizing it through a heuristic optimization algorithm, as Figure 2 shown, including:

[0073] S210. Obtain a design matrix according to the variable control points of the ship;

[0074] In the embodiment of the present invention, as Figure 3a shown, the ship is divided into fixed control points of the ship and moving control points of the bow. Specifically, a design matrix is obtained according to the variable control points of the ship, as Figure 3b shown, including:

[0075] S211. Determine at least some of the moving control points in the moving control points of the bow as the variable control points of the ship;

[0076] In the embodiment of the present invention, some of the moving control points in the moving control points of the bow that need to be deformed can be determined as the variable control points of the ship, and a stretching operation is performed on the variable control points of the ship.

[0077] S212. Determine the constraint range of the variable control points of the ship;

[0078] In the embodiment of the present invention, the constraint range of the variable control points of the ship is determined, that is, the stretching operation range of these variable control points is determined.

[0079] S213. Obtain a design matrix through stratified sampling by Latin hypercube sampling according to the constraint range of the variable control points of the ship.

[0080] A design matrix can be obtained through stratified sampling by Latin hypercube sampling within the constraint range.

[0081] As Figure 4 shown, it is a schematic diagram of Latin hypercube sampling provided by the present invention. As Figure 5 shown, specifically, it is a flow chart of Latin hypercube sampling, including:

[0082] S1: Divide each dimension into m non-overlapping intervals so that each interval has the same probability;

[0083] S2: Randomly select a point in each interval of each dimension;

[0084] S3: Then randomly select the points selected in S2 from each dimension and form a vector with them.

[0085] The design matrix can be determined through the above Latin hypercube sampling process.

[0086] S220. Obtain a deformed new hull line after deforming the hull line of the mother ship of the ship through a radial basis function;

[0087] It should be understood that in order to achieve the hull line transformation, it can be obtained by transforming it into a mathematical expression form through a radial basis function.

[0088] It should be noted here that the radial basis function is a scalar function symmetric along the radial direction, usually defined as the Euclidean distance between any point P in space and a certain center P i The mathematical expression is:

[0089] δ(||P - P i ||) i = 1, 2,..., n,

[0090] where P i represents the center of the function, ||P - P i || represents the Euclidean distance from P to P i δ(x) is a function with ||P - P i || as the independent variable. As Figure 6 shown, the expression of the δ(x) basis function can be divided into two categories, fully supported basis functions and compactly supported basis functions. The selection of the basis function needs to satisfy that the equation has a solution and the system of equations has stability.

[0091] Since the ship hull surface is constructed using a rectangular control point matrix and the movement of each control point is described using the corresponding basis function, the construction of the surface mathematical model can be realized by using the superposition of basis functions.

[0092] Specifically, after deforming the ship mother hull lines through radial basis functions, the hull parameters of the deformed new lines are obtained, such as Figure 7 shown, including:

[0093] S221. Parametrically express the ship mother hull lines, and parametrically express both the ship fixed control points and the bow moving control points;

[0094] S222. Perform variation processing on the design variables that are the ship variable control points among the bow moving control points according to the radial basis functions to obtain the deformed new lines.

[0095] It should be understood that in the embodiments of the present invention, the geometric shape is changed by moving the bow variable control points, thereby generating new lines.

[0096] S230. Perform iterative calculations based on the design matrix and the hull parameters of the deformed new lines to obtain hydrodynamic calculation results;

[0097] In the embodiments of the present invention, hydrodynamic calculation results can be obtained by performing iterative calculations based on the design matrix and the hull parameters of the deformed new lines.

[0098] Specifically, perform iterative calculations based on the design matrix and the hull parameters of the deformed new lines to obtain hydrodynamic calculation results, such as Figure 8 shown, including:

[0099] S231. Determine the hull parameters of the deformed new lines and the number of iterative steps, where the hull parameters include the ship length, draft, and beam;

[0100] S232. Perform iterative calculations on the design matrix according to the preset number of iterative steps to obtain hydrodynamic calculation results, and the hydrodynamic calculation results at least include the wave-making resistance of the hull and the fluid characteristics around the hull.

[0101] Figure 9 is the hydrodynamic calculation flow chart provided by the present invention, as Figure 9 shown:

[0102] First, determine the hull parameters, such as: ship length, draft, beam;

[0103] Then, determine that the index to be optimized is the wave-making resistance;

[0104] Again, according to the ShipFlow software tutorial specifications, write calculation scripts to achieve parametric operation;

[0105] Finally, call the ShipFlow software to automatically iterate the design matrix, and call the XPAN module (potential flow calculation and solution) of the ShipFlow software to obtain the wave-making resistance of the hull and the fluid characteristics around the hull.

[0106] S240. Train the neural network model based on the ship form parameters of the new line after deformation and the hydrodynamic calculation results to obtain a neural network model to be optimized.

[0107] In the embodiment of the present invention, specifically, the ship form parameters of the new line after deformation can be used as the input, and the hydrodynamic calculation results can be used as the output. By training the neural network model with multiple sets of such corresponding ship form parameters and hydrodynamic calculation results, a neural network model to be optimized can be obtained.

[0108] Specifically, train the neural network model based on the ship form parameters of the new line after deformation and the hydrodynamic calculation results to obtain a neural network model to be optimized, as Figure 10 shown, including:

[0109] S241. Determine the optimization target in the hydrodynamic calculation results, and the optimization target in the hydrodynamic calculation results at least includes the wave-making resistance of the hull and / or the fluid characteristics around the hull;

[0110] In the embodiment of the present invention, the optimization target in the hydrodynamic calculation results can specifically be the wave-making resistance of the hull.

[0111] S242. Train the neural network model based on the optimization target in the hydrodynamic calculation results and the ship form parameters of the new line after deformation to obtain a neural network model to be optimized.

[0112] In the embodiment of the present invention, taking the calculation of wave-making resistance realized by the commercial software to automatically iterate the design matrix as an example, the ship form parameters of the new line after deformation are used as the input, and the wave-making resistance coefficient is used as the output to train the neural network model to obtain a neural network model to be optimized.

[0113] Specifically, train the neural network model based on the optimization target in the hydrodynamic calculation results and the ship form parameters of the new line after deformation to obtain a neural network model to be optimized, including:

[0114] Screen the optimization target in the hydrodynamic calculation results according to the preset screening conditions to obtain effective data of the optimization target, where the preset screening conditions at least include prior knowledge and design specifications;

[0115] Train the neural network model based on the effective data of the optimization target and the ship form parameters of the new line after deformation to obtain a neural network model to be optimized.

[0116] It should be understood that still taking the wave-making resistance as the target to be optimized in the hydrodynamic calculation results, screening the hydrodynamic calculation results through prior knowledge, design specifications, etc., can define the wave-making resistance coefficients that do not meet the conditions as bad points and eliminate them, only retaining the valid data.

[0117] In the embodiments of the present invention, whether it is the valid data of the target to be optimized or the hull parameters of the new line after deformation, when training through the neural network model, the data all undergoes non-linear transformation from the input layer through multiple hidden layers and finally reaches the output layer for classification or regression operations. Using the backpropagation algorithm, calculate the gradient of the network parameters through the loss function, and update the parameters according to the gradient to minimize the loss function. Finally, obtain the neural network model to be optimized.

[0118] It should be noted that the multi-layer perceptron in the neural network model in the embodiments of the present invention may specifically include a three-layer model, as Figure 11 shown, the first layer is the input layer, the second layer is the hidden layer, and the third layer is the output layer.

[0119] The output of the hidden layer is f(wx + b), where w represents the weight and b represents the bias. The initial weight and bias are taken as random values between [0, 1]. The activation function f selects the sigmoid function:

[0120] The output of the output layer is: softmax(w2X1 + b2), where X1 represents the output of the hidden layer.

[0121] S250. Perform optimization calculation on the neural network model to be optimized according to the heuristic optimization algorithm to obtain a pre-trained ship hull line optimization model.

[0122] In the embodiments of the present invention, use the heuristic optimization algorithm to perform numerical optimization on the neural network model to be optimized. Figure 12 is the flowchart of the heuristic optimization algorithm, as Figure 12 shown:

[0123] First, encode: Use binary encoding.

[0124] Then, initialize the population: Use a random population to initialize the evolutionary algorithm.

[0125] Again, evaluate the fitness of the population individuals: Call the simulation process to calculate the wave-making resistance coefficient.

[0126] Final selection (tournament selection): Select k individuals from the population with equal probability. According to the fitness value of each individual, select the individual with the best fitness (the smallest error) as the parent. Repeat this process until the specified number of individuals is selected. If the individual with the strongest adaptability (the smallest total error) participates in the tournament, it will be selected with a 100% probability; if an ordinary individual x participates in the tournament, it must be more adaptable than the other k individuals to be selected. Without prior knowledge, x has the probability of being selected.

[0127] Crossover (simulated binary crossover): Specifically, it refers to the operation of replacing part of the structures of two parent individuals to generate new individuals, thereby improving the search ability. Before the crossover operation, the individuals in the population must also be paired. The currently commonly used pairing strategy is random pairing, that is, the individuals in the population are paired randomly in pairs, and the crossover operation is carried out between the paired individuals.

[0128] Simulated binary crossover generates offspring Q1 and Q2 through parent S a and S b :

[0129]

[0130] where S a (k) represents the kth parameter, and S b (k), Q1(k), Q2(k) are similar, and β k is a random number generated by the following probability density function:

[0131]

[0132] where η c represents an arbitrary non - negative real number.

[0133] Mutation (Gaussian mutation): Mutation specifically replaces some genes in the chromosome encoding string with other genes, which is an indispensable part of the genetic algorithm. Its purpose is to improve the local search ability of the genetic algorithm, maintain the diversity of the population, and prevent premature phenomena.

[0134] The method of Gaussian mutation is to generate a random number that follows a Gaussian distribution and replace the real - number value in the previous gene. The mathematical expectation of the random number generated by this algorithm is the real - number value of the current gene. Assume that a chromosome consists of two parts (x, σ), where the first component x represents a point in the search space and the second component represents the variance σ. Then the offspring individuals are generated as follows:

[0135] σ′ = σe N(0,Δσ) ,

[0136] x′ = x + N(0, Δσ′),

[0137] where N(0, Δσ′) represents a Gaussian function with a mean of 0 and a variance of σ′.

[0138] Repeat the iterative process until the termination condition or the optimal goal is reached.

[0139] In the embodiment of the present invention, optimizing the neural network to - be - optimized model according to the heuristic optimization algorithm to obtain a pre - trained ship hull form optimization model includes:

[0140] Determining design variables according to the variable control points of the ship, and determining target variables according to the optimization target in the hydrodynamic calculation results;

[0141] Determining the algorithm type of the heuristic optimization algorithm and the algorithm parameters corresponding to the algorithm type, where the algorithm type includes any one of genetic algorithm, simulated annealing method, and ant colony algorithm, and the algorithm parameters include population, crossover rate, and mutation rate;

[0142] Calling the neural network to - be - optimized model for optimization calculation according to the determined algorithm type, algorithm parameters, design variables, and target variables to obtain a pre - trained ship hull form optimization model.

[0143] Specifically, in the embodiment of the present invention, the heuristic optimization algorithm is used for optimization calculation to obtain the optimal hull form, Figure 13 which is a flowchart for optimizing the neural network to - be - optimized model using the heuristic optimization algorithm to obtain the optimal hull form, as Figure 13 shown:[[]]END

[0144] 1) Design variable setting: The variable control points to be stretched and deformed, that is, the aforementioned variable control points of the ship;

[0145] 2) Target variable setting: In the embodiment of the present invention, the wave - making resistance coefficient is taken as an example;

[0146] 3) Constraint condition setting: This optimization problem is an unconstrained optimization, so no constraint conditions are set;

[0147] 4) Select the optimization algorithm and perform high - order settings. The content of high - order settings includes, for example, population, crossover rate, mutation rate, etc.;

[0148] 5) Call the high - efficiency surrogate model for automatic calculation;

[0149] 6) Optimize to obtain the optimal hull form solution.

[0150] In summary, the ship optimization design method provided by the present invention can quickly generate a design matrix through Latin hypercube sampling, transform the parent ship form using radial basis function (RBF), call commercial software to iterate the design matrix and generate results, use a neural network algorithm to train the calculation results and generate a neural network model to be optimized, and can set the design variables, target variables, and constraint conditions by selecting and setting the optimization algorithm, and call the optimized process generated by training to carry out numerical optimization to generate a pre-trained ship form optimization model. The ship optimization design method provided by the present invention can save a large amount of time for ship designers in the initial ship design stage and improve the ship design efficiency.

[0151] In addition, the ship optimization design method provided by the present invention also has the following advantages:

[0152] (1) The sampling method is diverse and extensible.

[0153] When carrying out optimization design for different objectives, the selected design variables also vary. Therefore, when exploring the space of design variables, the sampling methods selected also vary. The present invention provides an extended interface for the sampling method, and the sampling method can be selected according to different problems to be solved.

[0154] (2) Neural network training model

[0155] Use a neural network to train the calculation results and the corresponding inputs to obtain a neural network model to be optimized. Optimization and search through the neural network model to be optimized can save a large amount of time under the condition of obtaining equivalent results.

[0156] (3) Heuristic optimization algorithm

[0157] In the embodiment of the present invention, an intelligent heuristic optimization algorithm is used to optimize the wave-making resistance coefficient, etc. For the non-linear problem of ship optimization problems with many design variables, intelligent optimization can be carried out to avoid falling into local optima.

[0158] As another embodiment of the present invention, a ship optimization design device 100 is provided for implementing the ship optimization design method described above. Among them, as Figure 14 shown, it includes:

[0159] A determination module 110, configured to determine ship hydrodynamic performance target information according to ship design objectives;

[0160] An obtaining module 120, configured to input the ship hydrodynamic performance target information into a pre-trained ship form optimization model to obtain ship form target parameters matching the ship hydrodynamic performance target information;

[0161] A design module 130, configured to complete ship design according to the ship form target parameters;

[0162] Among them, the pre-trained ship hull form optimization model is obtained by training a neural network model based on multiple groups of corresponding ship form parameters and hydrodynamic calculation results and optimizing it through a heuristic optimization algorithm.

[0163] The ship optimization design device provided by the present invention determines the ship hydrodynamic performance target information according to the ship design goal, and then obtains the ship hull form target parameters matching the ship hydrodynamic performance target information based on the pre-trained ship hull form optimization model. Finally, the ship design can be completed based on the ship hull form target parameters. Before this ship design, the neural network model is trained with multiple groups of corresponding ship form parameters and hydrodynamic calculation results, and then the pre-trained ship hull form optimization model is obtained through a heuristic optimization algorithm. The obtained pre-trained ship hull form optimization model can be directly used to obtain the ship hull form target parameters. This ship optimization design method can effectively improve the ship design efficiency, and since the pre-trained ship hull form optimization model is trained based on multiple groups of matching ship form parameters and hydrodynamic calculation results and optimized through a heuristic optimization algorithm, it can also improve the reliability of ship design.

[0164] For the specific working principle of the ship optimization design device of the present invention, reference can be made to the description of the ship optimization design method in the previous text, which will not be elaborated here.

[0165] As another embodiment of the present invention, a storage medium is provided, which is used to store computer instructions. When the computer instructions are loaded and executed by a processor, the ship optimization design method described above is implemented.

[0166] In an embodiment of the present invention, a non-transitory computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions can execute the ship optimization design method in any of the above method embodiments. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memories.

[0167] As another embodiment of the present invention, an electronic device is provided, which includes a memory and a processor. The memory and the processor are communicatively connected. The memory is used to store computer instructions, and the processor is used to load and execute the computer instructions to implement the ship optimization design method described above.

[0168] As Figure 15 shown, the electronic device 80 may include: at least one processor 81, such as a CPU (Central Processing Unit), at least one communication interface 83, a memory 84, and at least one communication bus 82. Among them, the communication bus 82 is used to implement the connection and communication between these components. Among them, the communication interface 83 may include a display screen and a keyboard. Optionally, the communication interface 83 may further include a standard wired interface and a wireless interface. The memory 84 may be a high-speed RAM memory (Random Access Memory), or a non-volatile memory, such as at least one disk memory. Optionally, the memory 84 may further be at least one storage device located far from the aforementioned processor 81. Among them, an application program is stored in the memory 84, and the processor 81 calls the program code stored in the memory 84 to execute any of the above method steps.

[0169] Among them, the communication bus 82 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus 82 may be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 13 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0170] Among them, the memory 84 may include a volatile memory, such as a random-access memory (RAM); the memory may also include a non-volatile memory, such as a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD); the memory 84 may further include a combination of the above types of memories.

[0171] Among them, the processor 81 can be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP.

[0172] Among them, the processor 81 may further include a hardware chip. The above-mentioned hardware chip can be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The above-mentioned PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0173] Optionally, the memory 84 is further configured to store program instructions. The processor 81 can call the program instructions to implement the ship optimization design method as shown in the embodiments of the present invention. Figure 1 Embodiments.

[0174] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principles of the present invention, and the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered within the protection scope of the present invention.

Claims

1. A ship optimization design method, characterized in that: include: Determine the target information of ship hydrodynamic performance according to the ship design objectives; Inputting the ship hydrodynamic performance target information into a pre-trained ship line optimization model to obtain ship line target parameters matching the ship hydrodynamic performance target information; Complete ship design according to the target ship line parameters; The pre-trained ship line optimization model is obtained by training a neural network model and optimizing it through a heuristic optimization algorithm according to multiple groups of corresponding ship type parameters and hydrodynamic calculation results; The pre-trained ship line optimization model is obtained by training a neural network model and optimizing it through a heuristic optimization algorithm according to multiple groups of corresponding ship type parameters and hydrodynamic calculation results, and includes: Obtaining a design matrix based on the ship's variable control points; The ship's mother ship type line is deformed by radial basis function to obtain a deformed new line; Perform iterative calculation based on the design matrix and the ship type parameters of the new type line after deformation to obtain a hydrodynamic calculation result; According to the ship type parameters of the deformed new line and the hydrodynamic calculation results, a neural network model is trained to obtain a neural network model to be optimized; The neural network model to be optimized is optimized by performing optimization calculation according to the heuristic optimization algorithm to obtain a pre-trained ship line optimization model.

2. The ship optimization design method according to claim 1, characterized in that: The design matrix is ​​obtained based on the ship's variable control points, including: Determining at least part of the bow movable control points as ship variable control points, wherein the ship is divided into ship fixed control points and bow movable control points; Determining the constraint range of the variable control point of the ship; The design matrix is ​​obtained by stratified sampling through Latin hypercube sampling according to the constraint range of the variable control points of the ship.

3. The ship optimization design method according to claim 1, characterized in that: The ship's mother ship type line is deformed by radial basis function to obtain the ship type parameters of the deformed new line, including: The mother ship type line of the ship is expressed parameterized, and the fixed control point of the ship and the bow movable control point are expressed parameterized; According to the radial basis function, the design variables of the bow moving control points as the variable control points of the ship are changed to obtain the deformed new lines.

4. The ship optimization design method according to claim 1, characterized in that: The hydrodynamic calculation results are obtained by iteratively calculating the design matrix and the ship type parameters of the deformed new line, including: Determine the ship type parameters and the number of iteration steps of the new type line after deformation, wherein the ship type parameters include ship length, depth and width; The design matrix is ​​iteratively calculated according to a preset number of iteration steps to obtain a hydrodynamic calculation result, which at least includes the wave-making resistance of the hull and the fluid characteristics around the hull.

5. The ship optimization design method according to claim 1, characterized in that: The neural network model is trained according to the ship type parameters of the deformed new line and the hydrodynamic calculation results to obtain the neural network model to be optimized, including: Determining an objective to be optimized in the hydrodynamic calculation result, wherein the objective to be optimized in the hydrodynamic calculation result at least includes the wave-making resistance of the hull and / or the fluid characteristics around the hull; The neural network model is trained according to the target to be optimized in the hydrodynamic calculation result and the ship type parameters of the new type line after deformation to obtain the neural network model to be optimized.

6. The ship optimization design method according to claim 5, characterized in that: The neural network model is trained according to the target to be optimized in the hydrodynamic calculation result and the ship type parameters of the new type line after deformation to obtain the neural network model to be optimized, including: Screening the target to be optimized in the hydrodynamic calculation results according to preset screening conditions to obtain valid data of the target to be optimized, wherein the preset screening conditions at least include prior knowledge and design specifications; The neural network model is trained according to the effective data of the target to be optimized and the ship type parameters of the deformed new line to obtain the neural network model to be optimized.

7. The ship optimization design method according to claim 1, characterized in that: The method further comprises: performing optimization calculation on the neural network model to be optimized according to the heuristic optimization algorithm to obtain a pre-trained ship line optimization model. Determining design variables according to the variable control points of the ship, and determining target variables according to the target to be optimized in the hydrodynamic calculation results; Determining the algorithm type of the heuristic optimization algorithm and algorithm parameters corresponding to the algorithm type, wherein the algorithm type includes any one of a genetic algorithm, a simulated annealing method and an ant colony algorithm, and the algorithm parameters include a population, a crossover rate and a mutation rate; The neural network model to be optimized is called according to the determined algorithm type, algorithm parameters, design variables and target variables to perform optimization calculations to obtain a pre-trained ship line optimization model.

8. A ship optimization design device, used to implement the ship optimization design method according to any one of claims 1 to 7, characterized in that: include: A determination module, used to determine target information of ship hydrodynamic performance according to ship design objectives; An acquisition module, used for inputting the ship hydrodynamic performance target information into a pre-trained ship line optimization model to obtain ship line target parameters matching the ship hydrodynamic performance target information; A design module, used to complete the ship design according to the target parameters of the ship line; The pre-trained ship line optimization model is obtained by training a neural network model based on multiple groups of corresponding ship type parameters and hydrodynamic calculation results and optimizing through a heuristic optimization algorithm.

9. A storage medium, characterized in that: Used to store computer instructions, which, when loaded and executed by a processor, implement the ship optimization design method described in any one of claims 1 to 7.

Images