Self-adaptive optimization method and system for blade parameters of rim propeller

Abstract

The invention relates to a rim propeller blade parameter adaptive optimization method, which comprises the following steps of: expressing an RDT blade through a characteristic curve to carry out parametric modeling to obtain at least more than two RDT blade models; s2, using a Sobo1 sequence algorithm to generate a plurality of sample schemes of the RDT blade model, and performing result collection on each sample scheme; s3, setting an objective function and constraint conditions, and training the verification model according to the plurality of sample schemes and the corresponding collection results to obtain a trained verification model; and S4, in combination with the verification model, an optimal RDT blade model is obtained by using a second-generation non-dominated sorting genetic algorithm, different optimal design schemes are verified through a CFD method, and performance differences between different optimal RDT blade models are compared.

Description

Technical Field

[0001] The invention relates to the technical field of blade design, and in particular to a method and system for adaptively optimizing parameters of a rim propeller blade. Background Art

[0002] At present, rim-driven thrusters (RDT) are widely used in ship propulsion systems due to their unique shaftless design. The main advantages of rim-driven thrusters include compact structure, reduced vibration and noise, and reduced cavitation. As the core component of rim-driven thrusters, the design quality of RDT blades directly affects the overall performance of the thruster. However, traditional blade design methods mostly rely on empirical formulas or fixed design processes, and it is difficult to fully consider the impact of different design parameters on blade performance. This results in the inability to find the optimal blade shape and parameter combination during the design process, which limits the performance improvement of the thruster.

[0003] Existing design methods often cannot quickly and accurately determine an efficient blade solution during the design phase, resulting in a long design cycle and high costs, and the final blade design may not fully meet actual needs. Therefore, a method that can automatically optimize blade parameters during the design process and improve design efficiency and blade performance is urgently needed. Summary of the invention

[0004] In view of the above-mentioned defects, the purpose of the present invention is to propose a method and system for adaptive optimization of rim thruster blade parameters, so as to solve the problem in the prior art that manual continuous testing is required to obtain optimized blade parameters, thereby greatly improving the efficiency of obtaining optimization solutions.

[0005] To achieve this purpose, the present invention adopts the following technical solution: a method for adaptively optimizing parameters of a rim propeller blade, comprising the following steps:

[0006] Step S1: performing parameterized modeling of the RDT blade by expressing the characteristic curve to obtain at least two or more RDT blade models;

[0007] Step S2: using the Sobo1 sequence algorithm to generate several sample solutions of the RDT blade model, and collecting results for each sample solution;

[0008] Step S3: setting the objective function and constraint conditions, training the verification model according to several sample schemes and corresponding collected results, and obtaining a trained verification model;

[0009] Step S4: Combined with the verification model, the second-generation non-dominated sorting genetic algorithm is used to obtain the optimal RDT blade model, and different optimal design schemes are verified by the CFD method to compare the performance differences between different optimal RDT blade models.

[0010] Preferably, the specific steps of step S1 are as follows:

[0011] Based on the setting requirements of the RDT blade, the corresponding characteristic curve is selected, and the parametric modeling of the RDT blade is completed by introducing the characteristic curve based on the NSRDC-NACA66-mod airfoil, wherein the characteristic curve includes a B-Spline curve and a F-Spline curve.

[0012] Preferably, the specific steps of step S2 are as follows:

[0013] Determine the parameter space for sampling;

[0014] Generate a series of sample points in a determined parameter space according to the Sobol sequence algorithm;

[0015] The open water efficiency η and cavitation initiation risk assessment index C are calculated by sample points. pmin / ρ, total torque coefficient 10K Q The mutual influence relationship between them; setting the mutual influence relationship as the result of collection.

[0016] Preferably, the objective function and constraint conditions in step S3 are as follows:

[0017] maxη

[0018] maxC pmin / ρ;

[0019]

[0020] Where η is the open water efficiency, C pmin is the minimum pressure coefficient, ρ is the cavitation number, K Q is the torque coefficient,

[0021] y(p i ) is the value of the control point of the blade design parameter characteristic curve, where p i represents the control point of the blade design parameter characteristic curve. The specific design parameters are chord length and thickness. K Qref Represents the torque coefficient reference value, y(d i ) is the value of the blade tip point where the wheel rim is connected, where d i Indicates the camber and pitch at the blade tip where the rim connects.

[0022] Preferably, the specific steps of training the verification model according to the collection results in step S3 are as follows:

[0023] Step S1: Use the sample solution as the input layer variable and the result as the output layer variable;

[0024] Step S2: Use the mean square error loss function to measure the difference between the predicted value and the actual value;

[0025] For each sample solution, the square of the difference between the predicted value and the actual value is calculated, and then the loss function value is averaged over all sample solutions. Where N is the number of sample solutions, is the predicted value of the nth sample, is the actual value of the nth sample;

[0026] Step S3: According to the loss function value, use the back propagation algorithm to update the connection weights and biases of the neural network; starting from the output layer, calculate the partial derivative of the loss function with respect to each connection weight and bias, and then adjust the connection weights and biases according to the size and direction of the partial derivatives so that the loss function value gradually decreases;

[0027] Step S4: repeat steps S2 to S3 until the loss function value is less than the loss threshold;

[0028] Step S5: Obtain the determination coefficient, and determine whether the determination coefficient is greater than the coefficient threshold. If it is less than, reconstruct the verification model; if it is greater than, output the verification model;

[0029] The determination coefficient is as follows:

[0030]

[0031] in Represents the mean of the true values.

[0032] Preferably, the use process of the second generation non-dominated sorting genetic algorithm in step S4 is as follows:

[0033] Step S41: first determine the population individuals, determine the initial individuals of the population according to the results and the verification model, and set the size of the initial population;

[0034] Step S42: for each individual in the population, the fitness value is calculated using the verification model according to the set objective function;

[0035] Step S43: sorting the individuals in the population by non-dominance, and classifying the individuals into different non-dominance levels according to the relationship between dominance and domination;

[0036] Step S44: According to the calculation results of non-dominated sorting and crowding distance, a part of individuals are selected as the parents of the next generation population; a crossover operation is performed on the selected parent individuals to generate new offspring individuals, and a mutation operation is performed on the newly generated offspring individuals.

[0037] Step S45: Repeat the above step S44 to continuously perform iterative optimization, and the iterative process continues until the number of iterations meets the number threshold.

[0038] Preferably, the specific CFD working mode parameters in step S4 include the following steps:

[0039] Get the operational procedures required for rim-thruster propeller simulation;

[0040] Record a series of operations in the simulation process as a .java file through recorded macro commands;

[0041] Combining the .sim source file and the .java file, calling the STAR CCM+ solver to batch process the design solutions;

[0042] Perform computational domain generation, meshing, and export of computational results to obtain simulation results of performance.

[0043] A rim propeller blade parameter adaptive optimization system, using the rim propeller blade parameter adaptive optimization method, including a modeling module, a sample generation module, a model training module and a verification module;

[0044] The modeling module is used to express the RDT blade through the characteristic curve for parameterized modeling, and obtain at least two or more RDT blade models;

[0045] The sample generation module is used to generate several sample solutions of the RDT blade model using the Sobol sequence algorithm, and collect results for each sample solution;

[0046] The model training module is used to set the objective function and constraint conditions, train the verification model according to several sample schemes and corresponding collected results, and obtain the trained verification model;

[0047] The verification module is used to: combine the verification model, use the second-generation non-dominated sorting genetic algorithm to obtain the optimal RDT blade model, and verify different optimal design schemes through the CFD method to compare the performance differences between different optimal RDT blade models.

[0048] One of the above technical solutions has the following advantages or beneficial effects: In the present invention, a large number of sample solutions can be quickly generated through parametric modeling and Sobol sequence algorithm, and the results can be collected and analyzed. This greatly improves the design efficiency and shortens the research and development cycle. The objective function and constraints are set to train the BP neural network model to obtain an approximate model, and then the second-generation non-dominated sorting genetic algorithm (NSGA-II) is used to search for the optimal design solution, thereby realizing the adaptive optimization of the blades under different working conditions. This method can be applied to different types of RDT blade designs, and can be expanded and improved according to actual needs. This helps to promote the continuous development and innovation of related technologies. In this way, the propulsion efficiency and stability of the rim thruster in different working environments can be significantly improved, thereby improving the overall performance and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of an embodiment of the method of the present invention.

[0050] Figure 2 It is a schematic diagram of the structure of an embodiment of the system of the present invention. DETAILED DESCRIPTION

[0051] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions from beginning to end. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0052] In the description of the embodiments of the present invention, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0053] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, the meaning of "plurality" is two or more. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0054] like Figures 1-2As shown, a method for adaptively optimizing parameters of a rim propeller blade comprises the following steps:

[0055] Step S1: performing parameterized modeling of the RDT blade by expressing the characteristic curve to obtain at least two or more RDT blade models;

[0056] Step S2: using the Sobo1 sequence algorithm to generate several sample solutions of the RDT blade model, and collecting results for each sample solution;

[0057] Step S3: setting the objective function and constraint conditions, training the verification model according to several sample schemes and corresponding collected results, and obtaining a trained verification model;

[0058] Step S4: Combined with the verification model, the second-generation non-dominated sorting genetic algorithm is used to obtain the optimal RDT blade model, and the different optimal design schemes are verified by the CFD method (Computational Fluid Dynamics, a numerical analysis method that uses computers to simulate fluid flow and heat transfer processes) to compare the performance differences between different optimal RDT blade models.

[0059] In the present invention, through parametric modeling and Sobol sequence algorithm, a large number of sample solutions can be quickly generated, and the results can be collected and analyzed. This greatly improves the design efficiency and shortens the research and development cycle. The objective function and constraints are set to train the BP neural network model to obtain an approximate model, and then the second-generation non-dominated sorting genetic algorithm (NSGA-II) is used to search for the optimal design solution, thereby realizing the adaptive optimization of the blades under different working conditions. This method can be applied to different types of RDT blade designs, and can be expanded and improved according to actual needs. This helps to promote the continuous development and innovation of related technologies. In this way, the propulsion efficiency and stability of the rim thruster in different working environments can be significantly improved, thereby improving the overall performance and reliability.

[0060] An example is used below to illustrate: first, the user can express the RDT blade through the characteristic curve for parameterized modeling, and then a plurality of different RDT blade models can be obtained, such as blade model 1, blade model 2, blade model 3, blade model n, etc.;

[0061] Then, the Sobo1 sequence algorithm can be used to generate the scheme of the RDT blade model. For example, in the first sample scheme of blade model 1, the length of the blade is X1 and the width is Y1. In the second sample scheme, the twist of the blade is set to X2 and the length is set to Y2. Multiple groups of different sample schemes can be generated by the Sobo1 sequence algorithm. At the same time, the corresponding result samples are collected. For example, in this optimization, if the open water efficiency needs to be optimized, the open water efficiency in different sample schemes is obtained. Then the corresponding adopted results are obtained.

[0062] Then the sample solution and the adopted results are input into the training model for training. After training, the training model can output the corresponding results according to the sample solution. Therefore, the optimal design solution for each RDT blade model can be found through the second-generation non-dominated sorting genetic algorithm. Finally, the performance of each optimal RDT blade model is compared to select the ideal blade design solution immediately.

[0063] Preferably, the specific steps of step S1 are as follows:

[0064] Based on the setting requirements of the RDT blade, the corresponding characteristic curve is selected, and the parametric modeling of the RDT blade is completed by introducing the characteristic curve based on the NSRDC-NACA66-mod airfoil, wherein the characteristic curve includes a B-Spline curve and a F-Spline curve.

[0065] The B-Spline curve has good local controllability and smoothness, and the F-Spline curve has advantages in fitting certain specific shapes. In order to further fit the real parameters for modeling, the actual geometry and design requirements of the blades are considered. If the geometry of the blade is more complex and requires high-precision fitting, you can choose the F-Spline curve that pays more attention to the curve's ability to fit complex shapes. If you need to emphasize that local control and smoothness have a greater impact on blade performance, you can choose the B-Spline curve. By introducing the characteristic curve to express the fully parametric modeling method of the RDT hydraulic component, the geometric characteristics of the blade can be described more accurately. Thereby, more realistic test results can be obtained in the subsequent CFD method verification.

[0066] In one embodiment of the present invention, the NSRDC-NACA66-mod airfoil is used for parametric modeling because the NSRDC-NACA66-mod airfoil is widely used and can better meet the design requirements of rim thruster blades. Although other types of airfoils may also have certain applicability, the NSRDC-NACA66-mod has abundant experimental data, strong application, and is suitable for most scenario requirements.

[0067] It is worth noting that when selecting the B-Spline curve, in order to prevent the default straight closed airfoil curve in the parametric modeling software from causing an uneven trailing edge connection, it is necessary to define the range of the blade in the circumferential (U direction) and radial (V direction) directions to extract the suction surface and pressure surface, and then use the Fillet command to fill the leading edge and following edge between the two surfaces. At the same time, the tip of the blade close to the hub is filled with Feature, and the Feature at the blade tip is compiled in a custom way.

[0068] Preferably, the specific steps of step S2 are as follows:

[0069] Determine the parameter space for sampling; that is, determine which blade parameters need to be sampled. These parameters may include the geometric dimensions and shape parameters of the blades;

[0070] A series of sample points are generated in a determined parameter space according to the Sobol sequence algorithm; it is worth mentioning that each sample point needs to correspond to a set of RDT blade models;

[0071] The open water efficiency η and cavitation initiation risk assessment index C are calculated by sample points. pmin / σ, total torque coefficient 10K Q The mutual influence relationship between them; setting the mutual influence relationship as the result of collection.

[0072] Preferably, the objective function and constraint conditions in step S3 are as follows:

[0073] maxη

[0074] maxC pmin / σ;

[0075]

[0076] Where η is the open water efficiency, C pmin is the minimum pressure coefficient, σ is the cavitation number, K Q is the torque coefficient,

[0077] y(p i ) is the value of the control point of the blade design parameter characteristic curve, where p i represents the control point of the blade design parameter characteristic curve. The specific design parameters are chord length and thickness. K Qref Represents the torque coefficient reference value, y(d i ) is the value of the blade tip point where the wheel rim is connected, where d i Indicates the camber and pitch at the blade tip where the rim connects.

[0078] Preferably, the specific steps of training the verification model according to the collection results in step S3 are as follows:

[0079] Step S1: Use the sample solution as the input layer variable and the result as the output layer variable;

[0080] In one embodiment of the present invention, the neural network of the training verification model includes the number of neurons in the input layer, the hidden layer and the output layer. The neural network consists of an input layer, two hidden layers and an output layer. The connection weights and biases of the neural network are initialized, and for each sample data, the design parameter values ​​of the input layer are transferred to the hidden layer through the connection weights. The output value of the hidden layer is then transferred to the output layer through the connection weights, and the input value of the output layer neuron is also calculated, and the predicted value of the output layer (predicted open water efficiency, cavitation initiation risk assessment index, total thrust coefficient and total torque coefficient) is obtained through the activation function.

[0081] Step S2: Use the mean square error loss function to measure the difference between the predicted value and the actual value;

[0082] For each sample solution, the square of the difference between the predicted value and the actual value is calculated, and then the loss function value is averaged over all sample solutions. Where N is the number of sample solutions, is the predicted value of the nth sample, is the actual value of the nth sample;

[0083] Step S3: According to the loss function value, use the back propagation algorithm to update the connection weights and biases of the neural network; starting from the output layer, calculate the partial derivative of the loss function with respect to each connection weight and bias, and then adjust the connection weights and biases according to the size and direction of the partial derivatives so that the loss function value gradually decreases;

[0084] Step S4: repeat steps S2 to S3 until the loss function value is less than the loss threshold;

[0085] Step S5: Obtain the determination coefficient, and determine whether the determination coefficient is greater than the coefficient threshold. If it is less than, reconstruct the verification model; if it is greater than, output the verification model;

[0086] The determination coefficient is as follows:

[0087]

[0088] in Represents the mean of the true values.

[0089] Preferably, the use process of the second generation non-dominated sorting genetic algorithm in step S4 is as follows:

[0090] Step S41: first determine the population individuals, determine the initial individuals of the population according to the results and the verification model, and set the size of the initial population;

[0091] Step S42: for each individual in the population, the fitness value is calculated using the verification model according to the set objective function;

[0092] The verification model quickly predicts the corresponding output variable values ​​(open water efficiency, cavitation initiation risk assessment index, etc.) according to the individual design parameters. These predicted values ​​are used as the basis for evaluating the individual performance, thereby determining the individual fitness value;

[0093] Step S43: Perform non-dominated sorting on the individuals in the population, and classify the individuals into different non-dominated levels according to the relationship between dominance and domination; in order to maintain the diversity of the population, calculate the crowding distance of each individual.

[0094] Step S44: According to the calculation results of non-dominated sorting and crowding distance, select a part of individuals as the parents of the next generation population; perform crossover operation on the selected parent individuals to generate new offspring individuals to increase the diversity of the population. Perform mutation operation on the newly generated offspring individuals to further increase the diversity and search ability of the population;

[0095] Step S45: Repeat the above step S44 to continuously perform iterative optimization, and the iterative process continues until the number of iterations meets the number threshold.

[0096] Finally, the best individual is selected from the final population as the optimal design solution for the rim thruster hydraulic assembly. The best individual is the individual with the best performance in the objective function space under the condition of satisfying the constraints, and its corresponding design parameters are the parameters of the optimal design solution.

[0097] Preferably, the specific CFD working mode parameters in step S4 include the following steps:

[0098] Get the operational procedures required for rim-thruster propeller simulation;

[0099] Record a series of operations in the simulation process as a .java file through recorded macro commands;

[0100] Combining the .sim source file and the .java file, calling the STAR CCM+ solver to batch process the design solutions;

[0101] Perform computational domain generation, meshing, and export of computational results to obtain simulation results of performance.

[0102] A rim propeller blade parameter adaptive optimization system, using the rim propeller blade parameter adaptive optimization method, including a modeling module, a sample generation module, a model training module and a verification module;

[0103] The modeling module is used to express the RDT blade through the characteristic curve to perform parameterized modeling, and obtain at least two or more RDT blade models;

[0104] The sample generation module is used to generate several sample solutions of the RDT blade model using the Sobol sequence algorithm, and collect results for each sample solution;

[0105] The model training module is used to set the objective function and constraint conditions, train the verification model according to several sample schemes and corresponding collected results, and obtain the trained verification model;

[0106] The verification module is used to: combine the verification model, use the second-generation non-dominated sorting genetic algorithm to obtain the optimal RDT blade model, and verify different optimal design schemes through the CFD method to compare the performance differences between different optimal RDT blade models.

[0107] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0108] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for adaptive optimization of rim propeller blade parameters, characterized in that: The steps include: Step S1: performing parameterized modeling of the RDT blade by expressing the characteristic curve to obtain at least two or more RDT blade models; Step S2: using the Sobo1 sequence algorithm to generate several sample solutions of the RDT blade model, and collecting results for each sample solution; Step S3: setting the objective function and constraint conditions, training the verification model according to several sample schemes and corresponding collected results, and obtaining a trained verification model; Step S4: Combined with the verification model, the second-generation non-dominated sorting genetic algorithm is used to obtain the optimal RDT blade model, and different optimal design schemes are verified by the CFD method to compare the performance differences between different optimal RDT blade models.

2. A method for adaptive optimization of rim propeller blade parameters according to claim 1, characterized in that: The specific steps of step S1 are as follows: Based on the setting requirements of the RDT blade, the corresponding characteristic curve is selected, and the parametric modeling of the RDT blade is completed by introducing the characteristic curve based on the NSRDC-NACA66-mod airfoil, wherein the characteristic curve includes a B-Spline curve and a F-Spline curve.

3. The method for adaptive optimization of rim propeller blade parameters according to claim 1, characterized in that: The specific steps of step S2 are as follows: Determine the parameter space for sampling; Generate a series of sample points in a determined parameter space according to the Sobol sequence algorithm; The open water efficiency η and cavitation initiation risk assessment index C are calculated by sample points. pmin / σ, total torque coefficient 10K Q The mutual influence relationship between them; setting the mutual influence relationship as the result of collection.

4. The method for adaptive optimization of rim propeller blade parameters according to claim 1, characterized in that: The objective function and constraint conditions in step S3 are as follows: maxη maxC pmin / σ; Where η is the open water efficiency, C pmin is the minimum pressure coefficient, ρ is the cavitation number, K Q is the torque coefficient, y(p i ) is the value of the control point of the blade design parameter characteristic curve, where p i represents the control point of the blade design parameter characteristic curve. The specific design parameters are chord length and thickness. K Qref Represents the torque coefficient reference value, y(d i ) is the value of the blade tip point where the wheel rim is connected, where d i Indicates the camber and pitch at the blade tip where the rim connects.

5. The method for adaptively optimizing blade parameters of a rim propeller according to claim 1, characterized in that: The specific steps of training the verification model according to the collection results in step S3 are as follows: Step S1: Use the sample solution as the input layer variable and the result as the output layer variable; Step S2: Use the mean square error loss function to measure the difference between the predicted value and the actual value; For each sample solution, the square of the difference between the predicted value and the actual value is calculated, and then the loss function value is averaged over all sample solutions. Where N is the number of sample solutions, is the predicted value of the nth sample, is the actual value of the nth sample; Step S3: According to the loss function value, use the back propagation algorithm to update the connection weights and biases of the neural network; starting from the output layer, calculate the partial derivative of the loss function with respect to each connection weight and bias, and then adjust the connection weights and biases according to the size and direction of the partial derivatives so that the loss function value gradually decreases; Step S4: repeat steps S2 to S3 until the loss function value is less than the loss threshold; Step S5: Obtain the determination coefficient, and determine whether the determination coefficient is greater than the coefficient threshold. If it is less than, reconstruct the verification model; if it is greater than, output the verification model; The determination coefficients are as follows: Where y represents the mean of the true values.

6. A method for adaptive optimization of rim propeller blade parameters according to claim 1, characterized in that: The specific process of using the second generation non-dominated sorting genetic algorithm in step S4 is as follows: Step S41: first determine the population individuals, determine the initial individuals of the population according to the results and the verification model, and set the size of the initial population; Step S42: for each individual in the population, the fitness value is calculated using the verification model according to the set objective function; Step S43: sorting the individuals in the population by non-dominance, and classifying the individuals into different non-dominance levels according to the relationship between dominance and domination; Step S44: According to the calculation results of non-dominated sorting and crowding distance, a part of individuals are selected as the parents of the next generation population; a crossover operation is performed on the selected parent individuals to generate new offspring individuals, and a mutation operation is performed on the newly generated offspring individuals. Step S45: Repeat the above step S44 to continuously perform iterative optimization, and the iterative process continues until the number of iterations meets the number threshold.

7. A method for adaptive optimization of rim propeller blade parameters according to claim 1, characterized in that: The specific CFD working mode parameters in step S4 include the following steps: Get the operational procedures required for rim-thruster propeller simulation; Record a series of operations in the simulation process as a .java file through recorded macro commands; Combining the .sim source file and the .java file, calling the STAR CCM+ solver to batch process the design solutions; Perform computational domain generation, meshing, and export of computational results to obtain simulation results of performance.

8. A rim propeller blade parameter adaptive optimization system, using a rim propeller blade parameter adaptive optimization method according to any one of claims 1 to 7, characterized in that: It includes modeling module, sample generation module, model training module and verification module; The modeling module is used to express the RDT blade through the characteristic curve for parameterized modeling, and obtain at least two or more RDT blade models; The sample generation module is used to generate several sample solutions of the RDT blade model using the Sobol sequence algorithm, and collect results for each sample solution; The model training module is used to set the objective function and constraint conditions, train the verification model according to several sample schemes and corresponding collected results, and obtain the trained verification model; The verification module is used to: combine the verification model, use the second-generation non-dominated sorting genetic algorithm to obtain the optimal RDT blade model, and verify different optimal design schemes through the CFD method to compare the performance differences between different optimal RDT blade models.

Images