CN-122020842-A - Multi-objective optimization design method and equipment for spray pipe layout and readable storage medium

CN122020842ACN 122020842 ACN122020842 ACN 122020842ACN-122020842-A

Abstract

The invention relates to the technical field of aerodynamic layout design and optimization of aircrafts, in particular to a spray pipe layout multi-target optimization design method, equipment and a readable storage medium; the method comprises the steps of obtaining a sample data set through numerical simulation, training a proxy model based on the sample set to obtain a prediction model of target response, optimizing the proxy model by adopting a multi-target optimization algorithm to obtain a Pareto optimal solution set of the spray pipe layout, performing accuracy verification on the optimal solution, and expanding the sample and retraining based on a self-adaptive point adding strategy if accuracy requirements are not met until accuracy is met. According to the invention, a proxy model is constructed to replace time-consuming high-fidelity simulation, so that the spray pipe layout optimization efficiency is remarkably improved, and a spray pipe layout scheme which is balanced and optimized among multiple targets such as control force, interference moment, applicability and stability can be rapidly obtained.

Inventors

ZHANG JIAYUE
MA JIKUI
LIU YUWEI
CAO NING
LIU YAOFENG

Assignees

中国航天空气动力技术研究院

Dates

Publication Date: 20260512
Application Date: 20251224

Claims (10)

1. The multi-target optimal design method for the spray pipe layout is characterized by comprising the following steps of: s1, determining optimized design variables and optimization targets, wherein the design variables are jet pipe layout parameters influencing jet flow control effects, and the optimization targets are at least two performance indexes related to jet flow control force and interference; S2, acquiring a sample data set, namely generating sample points by adopting an experimental design method based on the value range of the design variable, and carrying out numerical simulation on a spray pipe layout scheme corresponding to each sample point to acquire a corresponding performance response value of the optimization target to form the sample data set; S3, constructing and training a proxy model, namely training the proxy model for establishing a mapping relation between design variables and optimization targets based on the sample data set; s4, multi-objective optimization and verification, namely carrying out optimization solution on the trained agent model by utilizing a multi-objective optimization algorithm to obtain a Pareto optimal solution set of the spray pipe layout; and S5, judging and iterating, namely outputting a final optimization scheme if the verification precision meets the preset requirement, otherwise, selecting new sample points from the Pareto optimal solution set or the area with high model uncertainty according to the self-adaptive point adding strategy, carrying out numerical simulation, adding the sample data set, and returning to the step S3 to retrain the proxy model until the precision requirement is met.
2. The nozzle placement multi-objective optimization design method according to claim 1, wherein in step S1, the nozzle placement parameters include at least two of a nozzle installation position on an aircraft, an installation angle, a jet outlet area, and a jet total pressure.
3. The method according to claim 1, wherein in step S1, the optimization targets include at least two targets of maximizing jet control force, minimizing disturbance torque caused by jet, and improving applicability and stability of the reaction control system.
4. The method according to claim 1, wherein the numerical simulation in step 2 includes solving a three-dimensional compressible Navier-Stokes equation to obtain aerodynamic parameters of the jet flow disturbance flow field.
5. The method according to claim 1, wherein in step S3, the proxy model is one of a kriging model, a radial basis function model, a polynomial response surface model, and a support vector regression model.
6. The nozzle placement multi-objective optimization design method according to claim 1, wherein in step S4, the multi-objective optimization algorithm is a non-dominant ordered genetic algorithm, a multi-objective particle swarm algorithm, or a decomposition-based multi-objective evolutionary algorithm.
7. The method according to claim 1, wherein in step S5, the adaptive addition point strategy selects a new sample point based on a maximum prediction variance, a maximum expected improvement, or a maximum Pareto front uncertainty criterion.
8. The nozzle placement multi-objective optimization design method according to claim 1, wherein the method is applicable to jet disturbance control placement optimization design of hypersonic aircraft.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the nozzle placement multi-objective rapid optimization design method of any one of claims 1-8 when the program is executed by the processor.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when executed by a processor, implements the nozzle layout multi-objective rapid optimization design method as claimed in any one of claims 1 to 8.

Description

Multi-objective optimization design method and equipment for spray pipe layout and readable storage medium Technical Field The invention relates to the technical field of aerodynamic layout design and optimization of aircrafts, in particular to a spray pipe layout multi-objective optimization design method, equipment and a readable storage medium. Background The jet flow control technology is an active flow control technology for directly changing the flight track or the gesture of the aircraft through jet flow reaction force, and has the advantages of high response speed, no limitation of airspace and speed domain, high control precision and the like. Future development of hypersonic aircraft, whose flight range involves low to high altitudes, covers dense atmospheres up to a lean atmosphere. The aerodynamic force change range is large under different environments, and the numerical calculation simulation consumes longer time. The agent model is an approximation of a real objective function in the agent model-based optimization method, and an approximate optimal solution can be obtained quickly, so that the method is more suitable for processing complex engineering problems. Therefore, the construction of the agent model is particularly important for the rapid optimization design of the jet flow interference control layout. However, in the prior art, how to apply the proxy model to the nozzle layout efficiently and accurately, which involves the multi-objective optimization problem of strong nonlinearity and multi-physical field coupling, especially how to balance the optimization efficiency and precision, still lacks a systematic method. Therefore, a design method capable of rapidly and accurately completing multi-objective optimization of nozzle layout is needed. Disclosure of Invention The invention aims to provide a spray pipe layout multi-objective optimization design method, spray pipe layout multi-objective optimization design equipment and a readable storage medium, which can solve the technical problems. The invention provides a multi-target optimization design method for spray pipe layout, which comprises the following steps: s1, determining optimized design variables and optimization targets, wherein the design variables are jet pipe layout parameters influencing jet flow control effects, and the optimization targets are at least two performance indexes related to jet flow control force and interference; S2, acquiring a sample data set, namely generating sample points by adopting an experimental design method based on the value range of the design variable, and carrying out numerical simulation on a spray pipe layout scheme corresponding to each sample point to acquire a corresponding performance response value of the optimization target to form the sample data set; S3, constructing and training a proxy model, namely training the proxy model for establishing a mapping relation between design variables and optimization targets based on the sample data set; s4, multi-objective optimization and verification, namely carrying out optimization solution on the trained agent model by utilizing a multi-objective optimization algorithm to obtain a Pareto optimal solution set of the spray pipe layout; and S5, judging and iterating, namely outputting a final optimization scheme if the verification precision meets the preset requirement, otherwise, selecting new sample points from the Pareto optimal solution set or the area with high model uncertainty according to the self-adaptive point adding strategy, carrying out numerical simulation, adding the sample data set, and returning to the step S3 to retrain the proxy model until the precision requirement is met. Preferably, in step S1, the nozzle layout parameters include at least two of a mounting position, a mounting angle, a jet outlet area, and a jet total pressure of the nozzle on the aircraft. Preferably, in step S1, the optimization objectives include at least two objectives of maximizing jet control force, minimizing disturbance torque caused by the jet, and improving applicability and stability of the reaction control system. Preferably, the numerical simulation in the step 2 includes solving a three-dimensional compressible Navier-Stokes equation to obtain aerodynamic parameters of the jet flow disturbance flow field. Preferably, in step S3, the proxy model is one of a kriging model, a radial basis function model, a polynomial response surface model, or a support vector regression model. Preferably, in step S4, the multi-objective optimization algorithm is a non-dominant ordered genetic algorithm, a multi-objective particle swarm algorithm, or a decomposition-based multi-objective evolutionary algorithm. Preferably, in step S5, the adaptive dotting strategy performs selection of a new sample point based on a maximum prediction variance, a maximum expected improvement, or a maximum Pareto front uncertainty criterion. Preferably, the method is suitable