A design method and system for a semi-open fuel centrifugal pump impeller

Abstract

The present invention discloses a design method and system for a semi-open fuel centrifugal pump impeller, relating to the technical field of semi-open aviation fuel centrifugal pump impeller design. The method comprises the following steps: determining an optimization objective and optimization variables influencing the optimization objective; designing multiple sets of sample data corresponding to the optimization variables using a Latin square hypercube sampling method; establishing a semi-open fuel centrifugal pump impeller model based on the sample data corresponding to each set of optimization variables, performing numerical simulation, and establishing a database for training surrogate models; training the surrogate model using the database to obtain a prediction model; optimizing the optimization objective predicted by the prediction model using a multi-objective genetic optimization algorithm to obtain multiple sets of optimization variables that meet the optimization objective; performing numerical simulation on the multiple sets of optimization variables, and selecting a set of optimization variables corresponding to the simulation result with the best optimization objective from the multiple numerical simulation results as an optimization scheme for designing the semi-open fuel centrifugal pump impeller. The present invention can quickly design an optimization scheme for blades.

Description

Technical Field

[0001] The present invention relates to the technical field of semi-open aviation fuel centrifugal pump impeller design, and in particular to a design method and system for a semi-open fuel centrifugal pump impeller. Background Art

[0002] As a key energy source, centrifugal pumps are widely used in many industrial fields. Due to their wide flow range, strong boosting capabilities, and reliable and stable continuous operation, fuel centrifugal pumps serve as pre-boost pumps or afterburner pumps in aircraft engine fuel systems, supplying sufficient power to the combustion chamber. Their performance significantly impacts the engine's operating state and performance. As aircraft engines face increasingly harsh operating environments, the operating boundaries of aviation fuel centrifugal pumps are becoming more variable and the operating environment more severe, leading to more pronounced flow issues. Therefore, it is essential to develop design methods that better align with these flow patterns, and integrated impeller design is a key technology for this.

[0003] In the field of centrifugal pump impeller optimization design, researchers have used Bezier curves and linear functions to control the distribution and stacking variation of the blade profile, combined with an artificial fish swarm algorithm, on a Matlab platform for optimization. Researchers have conducted multi-objective optimization design and simulation research based on loss models and the sequential quadratic programming algorithm (SQP). Researchers have optimized the efficiency and cavitation performance of centrifugal pumps by constructing a mathematical optimization model with the goal of improving hydraulic efficiency and cavitation performance, while also using the elimination of the hump on the performance curve and ensuring that the model pump is not overloaded as constraints. Researchers have constructed a Kriging proxy model based on Latin hypercube sampling to optimize the design of marine double-suction centrifugal pump impellers. Researchers have used orthogonal experimental methods to optimize the design of deep-well pump impellers.

[0004] In summary, while existing technologies use Bezier curves to control parameters such as blade profiles, most require exporting blade coordinates to algorithmic software, such as Matlab, to implement blade control. This method is cumbersome when a large number of samples are required, resulting in time-consuming and computationally intensive early steps in computational fluid dynamics. Therefore, rapidly designing blade optimization solutions is a critical issue that needs to be addressed. Summary of the Invention

[0005] The embodiments of the present invention provide a design method and system for a semi-open fuel centrifugal pump impeller, which can solve the problem in the prior art that an optimization solution for blades cannot be quickly designed.

[0006] An embodiment of the present invention provides a method for designing a semi-open fuel centrifugal pump impeller, comprising the following steps:

[0007] The efficiency and head of a semi-open fuel centrifugal pump are selected as optimization objectives, and optimization variables that affect the optimization objectives are determined; wherein the optimization variables include: the profile line of the impeller hub and shroud on the axial plane projection, the blade thickness, and the blade skeleton line parameters;

[0008] The Latin square hypercube sampling method is used to design sample data corresponding to multiple groups of optimization variables. A semi-open fuel centrifugal pump impeller model is established based on the sample data corresponding to each group of optimization variables and numerical simulation is performed to obtain the value of the optimization target corresponding to the sample data corresponding to each group of optimization variables. The optimization target value corresponding to each group of optimization variables and the sample data corresponding to each group of optimization variables is used as centrifugal pump performance data, and a database is established.

[0009] A database is used to train a proxy model to predict the optimization target and obtain a prediction model. The optimization target predicted by the prediction model is optimized through a multi-objective genetic optimization algorithm to obtain multiple sets of optimization variables that meet the optimization target. Numerical simulation is performed on the multiple sets of optimization variables, and a set of optimization variables corresponding to the simulation result with the best optimization target is selected from multiple numerical simulation results as the optimization scheme for designing the impeller of a semi-open fuel centrifugal pump.

[0010] Furthermore, the design method of the semi-open fuel centrifugal pump impeller also includes: performing flow field analysis based on the optimization data in the optimization scheme and evaluating the optimization results.

[0011] Furthermore, the optimization target predicted by the prediction model is optimized by the multi-objective genetic optimization algorithm, and the specific steps include: using the multi-objective genetic optimization algorithm NSGA-II to optimize the optimization target until the preset termination condition is met to obtain an approximate Pareto solution set; and finding an optimization solution in the Pareto frontier of the approximate Pareto solution set.

[0012] Furthermore, the step of determining the optimization variables that affect the optimization target includes:

[0013] Perform reverse engineering on the three-dimensional model of the known semi-open fuel centrifugal pump impeller to obtain the impeller flow path;

[0014] According to the impeller flow channel, control points for controlling the shape and size of the blade are obtained;

[0015] Based on the control points, a model design tool is used to generate a parametric model similar to the known three-dimensional model of a semi-open fuel centrifugal pump impeller. The model design tool can parameterize the front and rear profiles, blade bone lines, and blade thickness using Bezier curves.

[0016] According to the parameterized model, sensitivity analysis is used to determine the degree of influence of the parameters in the model on the optimization target, and the parameters with the greatest influence on the optimization target are determined as optimization variables based on the degree of influence.

[0017] An embodiment of the present invention provides a design system for a semi-open fuel centrifugal pump impeller, comprising:

[0018] A parameter acquisition module is used to select the efficiency and head of the semi-open fuel centrifugal pump as optimization targets and determine optimization variables that affect the optimization targets; wherein the optimization variables include: the parameters of the impeller hub and shroud profile, blade thickness, and blade bone line on the axial plane projection;

[0019] A database establishment module is used to design sample data corresponding to multiple groups of optimization variables using the Latin square hypercube sampling method; establish a semi-open fuel centrifugal pump impeller model based on the sample data corresponding to each group of optimization variables and perform numerical simulation to obtain the optimization target value corresponding to the sample data corresponding to each group of optimization variables; use each group of optimization variables and the optimization target value corresponding to the sample data corresponding to each group of optimization variables as centrifugal pump performance data, and establish a database;

[0020] The solution acquisition module is used to use the database to train the proxy model to predict the optimization target and obtain the prediction model; optimize the optimization target predicted by the prediction model through a multi-objective genetic optimization algorithm to obtain multiple groups of optimization variables that meet the optimization target; perform numerical simulation on the multiple groups of optimization variables, and select a group of optimization variables corresponding to the simulation result with the best optimization target from multiple numerical simulation results as the optimization solution for designing the impeller of a semi-open fuel centrifugal pump.

[0021] The embodiment of the present invention provides a design method and system for a semi-open fuel centrifugal pump impeller. Compared with the prior art, the method and system have the following advantages:

[0022] Latin hypercube sampling is performed on the optimization variables, and a semi-open fuel centrifugal pump impeller model is established based on the sampling results, so as to obtain the optimization target value through simulation in the semi-open fuel centrifugal pump impeller model, and the optimization target value corresponding to each group of optimization variables and the sample data corresponding to each group of optimization variables is used as the centrifugal pump performance data, and a database is established. The optimization target is predicted by the proxy model trained by the database to obtain a prediction model, and the optimization target predicted by the prediction model is optimized by a multi-objective genetic optimization algorithm to obtain multiple groups of optimization variables that meet the optimization target; numerical simulation is performed on the multiple groups of optimization variables, and a group of optimization variables corresponding to the simulation result with the best optimization target is selected from multiple numerical simulation results as the optimization scheme for designing the semi-open fuel centrifugal pump impeller. This method can directly obtain the optimization target through a group of optimization variables, and realizes the rapid design of the blade optimization scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 An integrated blade model after reverse modeling of a design method for a semi-open fuel centrifugal pump impeller provided by an embodiment of the present invention;

[0024] Figure 2 The impeller hub and wheel cover control points of a semi-open fuel centrifugal pump impeller design method provided by an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of blade skeleton lines and related parameters of a design method for a semi-open fuel centrifugal pump impeller provided by an embodiment of the present invention;

[0026] Figure 4 Blade thickness control points and coordinates of a design method for a semi-open fuel centrifugal pump impeller provided by an embodiment of the present invention;

[0027] Figure 5 An impeller optimization workflow diagram of a design method for a semi-open fuel centrifugal pump impeller provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] To make the above-mentioned objects, features, and advantages of the present invention more readily apparent, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. The following description sets forth numerous specific details to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art may make similar modifications without departing from the scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0029] An embodiment of the present invention provides a method for designing a semi-open fuel centrifugal pump impeller, comprising the following steps:

[0030] Step 1: Select the efficiency and head of the semi-open fuel centrifugal pump as the optimization objectives, and determine the optimization variables that affect the optimization objectives.

[0031] Step 2: Use the Latin square hypercube sampling method to design sample data corresponding to multiple groups of optimization variables; establish a semi-open fuel centrifugal pump impeller model based on the sample data corresponding to each group of optimization variables and perform numerical simulation to obtain the optimization target value corresponding to the sample data corresponding to each group of optimization variables; use each group of optimization variables and the optimization target value corresponding to the sample data corresponding to each group of optimization variables as centrifugal pump performance data, and establish a database.

[0032] Step 3: Use the database to train a surrogate model to predict the optimization objective and obtain a prediction model. Surrogate models include orthogonal polynomial models, polynomial response surface models, radial basis function models, and Kriging models. A multi-objective genetic optimization algorithm is used to optimize the optimization objective predicted by the prediction model, obtaining multiple sets of optimization variables that meet the optimization objective. Numerical simulations are then performed on these multiple sets of optimization variables. From the multiple numerical simulation results, the set of optimization variables corresponding to the simulation result with the optimal optimization objective is selected as the optimization solution for the design of the semi-open fuel centrifugal pump impeller.

[0033] The solutions corresponding to the above steps are as follows:

[0034] (1) Select variables according to the optimization objectives.

[0035] After initially selecting the target optimization parameters based on the design requirements of a semi-open centrifugal pump, a parametric blade model was established. Through sensitivity analysis, the geometric parameters with the greatest impact on the optimization target were selected as the final optimization variables. The optimization target is the efficiency and head of the centrifugal pump at the design flow rate. The optimization variables include the profile parameters of the impeller hub and shroud on the axial projection, the blade thickness, and the blade skeleton parameters.

[0036] After the parametric model was established, it was imported into the meshing software Turbogrid to generate a high-quality hexahedral mesh and perform a mesh independence test. Because parametric optimization requires the calculation of a large number of experimental design models to ensure the accuracy of the results, a mesh independence test was performed before the calculations to minimize the number of meshes while ensuring the required accuracy. This allowed the selection of an appropriate mesh size as the meshing standard for subsequent parametric optimization calculations.

[0037] (2) Experimental design

[0038] The Latin square hypercube sampling design experimental method was used to establish experimental samples for each optimization variable, and these data were integrated as a database for training the proxy model. According to the geometric parameters given by each experimental scheme in the database, the parameters of the original parametric model of the aviation fuel centrifugal pump were modified in the impeller design software to generate the corresponding aviation fuel centrifugal pump model. Numerical simulations were performed on the aviation fuel centrifugal pump model corresponding to each experimental scheme to obtain the optimization target values of different samples and the simulation results of all experimental schemes.

[0039] (3) Build an agent model.

[0040] Use different surrogate model algorithms to establish the relationship between input variables and output variables, verify the accuracy of each surrogate model, and select the one with the best performance as the prediction model.

[0041] (4) Search for multi-objective optimization solutions.

[0042] An optimized solution is obtained by optimizing the optimization objectives using a multi-objective genetic optimization algorithm. This multi-objective optimization algorithm continues to iterate until a pre-set termination criterion is met, indicating that the solution quality meets a predetermined standard. Once the termination criterion is met, the individuals retained from the final environmental selection will form an approximate Pareto solution set. Designers should search for an appropriate optimization solution within the Pareto frontier based on their actual needs.

[0043] (5) Confirm the optimization results.

[0044] Numerical simulations were performed on the optimized solution. The optimization algorithm derived a combination of performance parameters, taking into account both hydraulic efficiency and head. The optimal combination of hydraulic efficiency and head was selected, and the combination with the highest combined hydraulic efficiency and head was chosen as the final optimized solution. A flow field analysis was conducted on the optimally designed aviation fuel centrifugal pump, and performance before and after optimization was compared to confirm the optimization results.

[0045] In part (1), the optimization target is first determined based on actual needs. In this paper, the efficiency and head of the design flow rate are simultaneously used as optimization targets. First, the design software is used to perform reverse processing on the three-dimensional impeller model of the centrifugal pump. The impeller parameterization tool is imported to parameterize the profile parameters, blade thickness, and blade skeleton of the impeller hub and shroud on the axial projection.

[0046] In part (2), Latin square hypercube sampling design is used to sample the optimization variables. This method has uniform sampling and good orthogonality. Different centrifugal pump impeller models are established based on the sampling data, and centrifugal pumps equipped with different centrifugal pump impellers are simulated and calculated to obtain centrifugal pump performance data as the test sample point database.

[0047] In part (3), different surrogate model algorithms are used to establish the relationship between input variables and output variables. Several surrogate models are compared at the theoretical level. Each method is used to establish the relationship between input variables and output variables in turn, and the model with the best performance is selected as the prediction model.

[0048] In part (4), the optimization solution is obtained by using a multi-objective genetic optimization algorithm to optimize the optimization objectives. The present invention is a multi-objective optimization algorithm, and the algorithm will continue to iterate until the preset termination condition is met, that is, the quality of the solution reaches the established standard. When the termination condition is met, the individuals retained in the last environmental selection will constitute an approximate Pareto solution set. Designers should find a suitable optimization solution in the Pareto frontier according to actual needs.

[0049] In summary, the integrated impeller optimization method proposed in the present invention greatly reduces the optimization threshold. The impeller parametric design method based on Bezier curve can effectively reduce the number of tests in the development process of centrifugal pumps. The parametric curve is used to achieve rapid control of the blade profile, providing a rich data basis for optimization design, which can effectively improve the design speed and accuracy.

[0050] An embodiment of the present invention provides a design system for a semi-open fuel centrifugal pump impeller, comprising:

[0051] The parameter acquisition module is used to select the efficiency and head of the semi-open fuel centrifugal pump as the optimization target and determine the optimization variables that affect the optimization target; wherein the optimization variables include: the parameters of the impeller hub and wheel cover on the axial plane projection, blade thickness and blade bone line.

[0052] The database establishment module is used to design sample data corresponding to multiple groups of optimization variables using the Latin square hypercube sampling method; establish a semi-open fuel centrifugal pump impeller model based on the sample data corresponding to each group of optimization variables and perform numerical simulation to obtain the optimization target value corresponding to the sample data corresponding to each group of optimization variables; use each group of optimization variables and the optimization target value corresponding to the sample data corresponding to each group of optimization variables as centrifugal pump performance data, and establish a database.

[0053] The solution acquisition module is used to use the database to train the proxy model to predict the optimization target and obtain the prediction model; optimize the optimization target predicted by the prediction model through a multi-objective genetic optimization algorithm to obtain multiple groups of optimization variables that meet the optimization target; perform numerical simulation on the multiple groups of optimization variables, and select a group of optimization variables corresponding to the simulation result with the best optimization target from multiple numerical simulation results as the optimization solution for designing the impeller of a semi-open fuel centrifugal pump.

[0054] A specific embodiment is as follows:

[0055] Based on the design requirements of the semi-open centrifugal pump, select the target parameters for optimization. The optimization target refers to the efficiency and head of the centrifugal pump's design flow rate. The optimization targets, determined based on the requirements, include the impeller hub and shroud profile, blade thickness, and blade bone line parameters on the axial projection. Then, make a preliminary selection of the geometric parameters to be optimized and their ranges.

[0056] The BladeEditor function of the enhanced blade geometry modeling of the 3D modeling module DesignModeler in ANSYS Workbench is used to perform reverse processing on the 3D impeller model of the centrifugal pump. Figure 1 As shown, the Flowpath impeller flow channel is first generated, and the control points ExportPoints of the long and short blades are regenerated to complete the reverse design.

[0057] The interactive turbomachinery blade design tool BladeGen is imported to implement Bezier curves for the line parameters, blade thickness, and blade bone line parameters of the impeller hub and shroud on the axial projection diagram, completing the generation of the parametric model.

[0058] The wheel cover contour line takes 3 control points, and the impeller hub contour segment takes 3 control points, such as Figure 2 As shown in the figure, all of them are determined by five-point quadratic Bezier curves, so the impeller hub and wheel cover have a total of 12 optimization parameters.

[0059] Select the six control points corresponding to the circular angles of the two streamlines at span=0 and span=1, such as Figure 3 As shown, it is determined by a five-point quadratic Bezier curve, i.e., 12 optimization parameters.

[0060] Select the streamline at span=0.5 to control the thickness of the impeller blade, such as Figure 4 As shown, it is determined by a five-point quartic Bezier curve with 6 optimization parameters.

[0061] A total of 30 optimization parameters were initially determined.

[0062] Import the Bladegen model into Turbogrid to automatically generate a high-quality hexahedral mesh. Meshes for other components (volute, inlet section) are generated using Meshing. Import the generated mesh into CFX or Fluent numerical simulation software to perform transient numerical simulations of the centrifugal pump under rated operating conditions. The simulation results provide the efficiency and head under these conditions.

[0063] On the basis of establishing the blade parameter model, the parameters with the greatest impact on the optimization target are selected as optimization variables through sensitivity calculation; the model is optimized by the parameter optimization software OptiSLang installed on the simulation software Workbench platform. The complete workflow is as follows: Figure 5 Based on the sensitivity analysis, the optimal response design parameters are obtained as the final optimization parameters. After the research object is converted into a mathematical model, it is assumed that the model has n parameters and have a functional relationship y = f ( x 1, x 2, ..., x n ), in addition to ensuring x i Assuming all other parameters remain constant, the parameter is allowed to vary within its possible range. The basic idea is to calculate the sensitivity coefficient of a parameter by varying its value one by one and observing the changes in the system output. Based on the calculated sensitivity coefficient, it is possible to determine which input parameters have a greater impact on the system output, thus providing a basis for subsequent optimization, adjustment, or decision-making.

[0064] Using the experimental design method provided by the Design Exploration module in ANSYS Workbench, in the Toolbox of Workbench, select Sensitivity under OptiSLang to enter the parametric design, and select the sampling method in DOE. For this optimization method, the Latin square hypercube sampling design test method is used to generate sample data for each optimization variable according to the experimental design. The number of samples is generally 2 ( n +1)( n +2) group ( n is the final number of optimized parameters after sensitivity analysis), different centrifugal pump impeller models are established according to the sampling data, and centrifugal pumps equipped with different centrifugal pump impellers are simulated and calculated to obtain centrifugal pump performance data as a database.

[0065] A surrogate model is used to predict the target optimization parameters of a centrifugal pump equipped with different impellers. First, different surrogate models are trained using a database. A variety of surrogate models provided by the ANSYS OptiSLang platform are used to meet different research needs. These include the Orthogonal Polynomial Model (OPM), the Polynomial Response Surface Model (RSM), the Radial Basis Function (RBF), and the Kriging model. In engineering research, in order to ensure the accuracy of the surrogate model and effectively control the time cost, RSMs of order 1 to 4 can be selected as surrogate models. This is because low-order RSMs can better capture system behavior and have relatively low computational complexity, making them suitable for rapid evaluation of design solutions. RBFs can flexibly handle nonlinear relationships and maintain good prediction performance when the number of samples is limited. After comparing the performance of each surrogate model, the present invention selects the RSM model with better performance as the final prediction model.

[0066] The NSGA-II algorithm is used to perform a multi-objective optimization of head and efficiency under the design conditions. The optimization objective is then optimized using a multi-objective genetic optimization algorithm to obtain an optimized solution. This multi-objective optimization algorithm continues to iterate until a preset termination condition is met, indicating that the solution quality meets a predetermined standard. Once the termination condition is met, the individuals retained from the final environmental selection will form an approximate Pareto solution set. Designers should search for an appropriate optimization solution within the Pareto front based on their actual needs.

[0067] Numerical simulations were performed on the optimized solution. The optimization algorithm derived a combination of performance parameters, taking into account both hydraulic efficiency and head. The combination with the highest hydraulic efficiency was selected as the final optimized solution. After obtaining the optimized solution, the resulting optimized data was substituted into the original model and simulated. The accuracy of the results was verified by comparing the proxy model and simulation results. A flow field analysis was performed on the optimal fuel centrifugal pump design, and performance before and after optimization was compared to confirm the optimization results.

[0068] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be determined by the appended claims.

Claims

1. A design method for a semi-open fuel centrifugal pump impeller, characterized in that: The following steps are involved: The efficiency and head of a semi-open fuel centrifugal pump are selected as optimization objectives, and optimization variables that affect the optimization objectives are determined; wherein the optimization variables include: the profile line of the impeller hub and shroud on the axial plane projection, the blade thickness, and the blade skeleton line parameters; The Latin square hypercube sampling method is used to design sample data corresponding to multiple groups of optimization variables. A semi-open fuel centrifugal pump impeller model is established based on the sample data corresponding to each group of optimization variables and numerical simulation is performed to obtain the value of the optimization target corresponding to the sample data corresponding to each group of optimization variables. The optimization target value corresponding to each group of optimization variables and the sample data corresponding to each group of optimization variables is used as centrifugal pump performance data, and a database is established. Use the database to train the proxy model to obtain a prediction model; use a multi-objective genetic optimization algorithm to optimize the optimization target predicted by the prediction model and obtain multiple sets of optimization variables that meet the optimization target; Numerical simulations are performed on multiple sets of optimization variables. From the multiple numerical simulation results, a set of optimization variables corresponding to the simulation result with the best optimization objective is selected as the optimization scheme for designing the impeller of a semi-open fuel centrifugal pump. The specific steps of determining the optimization variables that affect the optimization target include: Perform reverse engineering on the known three-dimensional model of the semi-open fuel centrifugal pump impeller to obtain the impeller flow path; According to the impeller flow channel, control points for controlling the shape and size of the blade are obtained; Based on the control points, a model design tool is used to generate a parametric model similar to the known three-dimensional model of a semi-open fuel centrifugal pump impeller. The model design tool can parameterize the front and rear profiles, blade bone lines, and blade thickness using Bezier curves. According to the parameterized model, sensitivity analysis is used to determine the degree of influence of the parameters in the model on the optimization target, and the parameters with the greatest influence on the optimization target are determined as optimization variables based on the degree of influence.

2. The design method of a semi-open fuel centrifugal pump impeller according to claim 1, characterized in that: The design method of the semi-open fuel centrifugal pump impeller also includes: Based on the optimization data in the optimization plan, flow field analysis is performed to evaluate the optimization results.

3. The design method of a semi-open fuel centrifugal pump impeller according to claim 1, characterized in that: The optimization target predicted by the prediction model is optimized by the multi-objective genetic optimization algorithm, and the specific steps include: The multi-objective genetic optimization algorithm NSGA-II is used to optimize the optimization objectives until the preset termination conditions are met and an approximate Pareto solution set is obtained; Find the optimal solution in the Pareto front of the approximate Pareto solution set.

4. A design system for a semi-open fuel centrifugal pump impeller, characterized in that: include: A parameter acquisition module is used to select the efficiency and head of the semi-open fuel centrifugal pump as optimization targets and determine optimization variables that affect the optimization targets; wherein the optimization variables include: the parameters of the impeller hub and shroud profile, blade thickness, and blade bone line on the axial plane projection; A database establishment module is used to design sample data corresponding to multiple groups of optimization variables using the Latin square hypercube sampling method; establish a semi-open fuel centrifugal pump impeller model based on the sample data corresponding to each group of optimization variables and perform numerical simulation to obtain the value of the optimization target corresponding to the sample data corresponding to each group of optimization variables; use the optimization target value corresponding to each group of optimization variables and the sample data corresponding to each group of optimization variables as centrifugal pump performance data, and establish a database; The solution acquisition module is used to use the database to train the proxy model to predict the optimization target and obtain the prediction model; optimize the optimization target predicted by the prediction model using a multi-objective genetic optimization algorithm to obtain multiple sets of optimization variables that meet the optimization target; perform numerical simulation on the multiple sets of optimization variables, and select a set of optimization variables corresponding to the simulation result with the best optimization target from the multiple numerical simulation results as the optimization solution for designing the impeller of the semi-open fuel centrifugal pump; The specific steps of determining the optimization variables that affect the optimization target include: Perform reverse engineering on the known three-dimensional model of the semi-open fuel centrifugal pump impeller to obtain the impeller flow path; According to the impeller flow channel, control points for controlling the shape and size of the blade are obtained; Based on the control points, a model design tool is used to generate a parametric model similar to the known three-dimensional model of a semi-open fuel centrifugal pump impeller. The model design tool can parameterize the front and rear profiles, blade bone lines, and blade thickness using Bezier curves. According to the parameterized model, sensitivity analysis is used to determine the degree of influence of the parameters in the model on the optimization target, and the parameters with the greatest influence on the optimization target are determined as optimization variables based on the degree of influence.

Images