Aerodynamic shape optimization method based on neural network integration

Abstract

The invention provides an aerodynamic shape optimization method based on neural network integration, and aims to solve the aircraft aerodynamic shape optimization problem. The method includes the steps that firstly, multiple aerodynamic shapes are constructed as samples according to different aerodynamic shape parameters, and the objective function of each sample is acquired by the adoption of a numeric analysis method; based on sample data, an objective function approximate model is constructed by the adoption of a neural network integration method; compared with a numeric analysis method, the approximate model is low in accuracy but much smaller in calculated amount; the objective function approximate model and a direct search algorithm are combined for optimization search, and the numeric analysis method or the approximate model is called based on a certain strategy in the search process to acquire the objective functions until the optimal aerodynamic shape is acquired. According to the method, on the premise that the optimization effect is guaranteed, times of numeric analysis can be effectively reduced in the optimization process, optimization design efficiency and quality can be improved, and the method is very suitable for the aircraft aerodynamic shape optimization design and relevant engineering problems.

Description

technical field [0001] The invention relates to the field of aircraft aerodynamic shape design. Background technique [0002] The traditional aerodynamic design method relies on the designer's experience and analysis ability to select the shape, and then uses aerodynamic analysis or wind tunnel experiments to evaluate and further modify the design scheme, so as to obtain a satisfactory aerodynamic design scheme repeatedly. When the shape is complex or there are many constraints, the entire optimization process is not easy to grasp, and it is very difficult to find the optimal shape that satisfies different constraints. [0003] The basic idea of ​​the numerical optimization method is to combine the optimization algorithm with the aerodynamic analysis software, transform the aerodynamic optimization design problem into a mathematical optimization problem, and rely on the numerical analysis method to solve the aerodynamic shape optimization design problem. Compared with the t...

AI Technical Summary

Problems solved by technology

Since it is time-consuming to solve the RANS equation numerically, and the aerodynamic characteristics analysis needs to be performed many times in the optimization, the numerical solution of the RANS equation as the aerodynamic analysis method will bring a large amount of calculation, and it is difficult to use it in engineering design

Images