Intelligent Method In The Application Of Civil Aircraft Cruise Configuration Aerodynamic Optimization

Aerodynamic optimization of civil aircraft is the basis of model development, and has a huge impact on their future competitiveness in the market. With the development of computer technology, database technology, and optimization methods in recent years, experts propose a new direction of development, that is, systematic, efficient and intelligent.In this paper, to solve the defects in the common aerodynamic optimization process, a new intelligent aerodynamic optimization design system is proposed. This system proposes improvements from the four aspects of the shape parametric approach, mesh generation, aerodynamic optimization methods and design experience accumulation to reflect systematic, efficient and intelligent, which agree the develop trend of civil aircraft aerodynamic optimization. By a neural network expert database, preservation and arrangement of experience in the design can be done; A improved particle swarm optimization method is used to enhance the optimization capabilities; Flow field calculation supplemented by BP neural network method is used in the optimization process, to enhance the optimization efficiency further.This paper focuses on the aerodynamic optimization of the cruise configuration of civilian airliner and main research issues and achievements are as follows:(1) Derive cubic non-uniform B-spline parametric approach for airfoil and wing, describe the aerodynamic shape accurately with small amount of parameters to reduce the number of optimization variables;(2) Generate the surface grid and flow field grid quickly using B-spline method and the Delaunay mesh morphing technology and increase the mesh generation speed without sacrificing grid quality;(3) Use the improved particle swarm optimization method for global aerodynamic optimization, and reduce the probability into a local minim. In the optimization process, use BP neural network to predict the aerodynamic performance and reduce the time spent on the CFD calculation and improve the optimization efficiency;(4) Establish a database of experts to save good optimization results, and provide convenience for the subsequent optimization of the design work. Use the SOM neural network analysis for the expert database can quickly search for the shape close to the design goals and enhance the ability of aerodynamic design.On the basis of the above work, for airfoil and wing-body combination, according to different optimization objectives, we finally get satisfactory optimization results. The results showed that the technique of CNUBS curves and surfaces can easily use less control vertex to describe the shape of the airfoil and wing more accurately, on the basis of which multi-objective aerodynamic optimization efficiency has improved. The design result has a good aerodynamic performance, which indicates the feasibility of above methods, and have a relatively high efficiency.