A Review of Aircraft Aerodynamic and Structural Optimization Method Based on Artificial Intelligence

Abstract

With aviation technology continuous development, aircraft design is going toward faster, lighter and more intelligent direction. Aerodynamic shape characteristic and structural stiffness are two very important factor that show aircraft overall performance, so optimize aerodynamic shape and structural stiffness always is hot research direction in aviation system engineering field. Traditional aerodynamic-structure optimization design method usually need consume many CFD and FEA large numerical simulation, and need bigger design freedom and higher design efficiency, hard to fit modern complex aviation system design process. These years, machine learning, deep learning, reinforcement learning and evolutionary algorithm etc. artificial intelligence technology develop very rapid, bring completely new idea for aircraft optimization design, can on data-driven base, through machine learning, deep learning, reinforcement learning and evolutionary algorithm etc. way complete aircraft aerodynamic characteristic prediction, aircraft structural response modeling, based on aircraft aerodynamic and structure multi-objective optimization model global optimization etc. task, thus can effective reduce aircraft design high calculation cost and improve design accuracy. This paper review machine learning, deep learning, reinforcement learning and evolutionary algorithm etc. artificial intelligence technology in aviation aircraft aerodynamic and structural optimization design development and future use technical characteristic, key introduce surrogate model building, through deep neural network do aircraft flow field prediction and shape reconstruction, based on reinforcement learning adaptive control and multi-objective optimization, evolutionary algorithm in aircraft lightweight design application etc. research method, discuss AI technology in aviation aircraft aerodynamic-structure integrated optimization design and multidisciplinary collaborative design application prospect and challenge problem, like data limited, model transparency not enough and multi-scale effect coupling problem etc. At last, put forward artificial intelligence apply to aircraft design further development direction, like physics guide neural network, data scarce learning, cloud intelligent calculation and base on artificial intelligence automation design platform, this paper give research summary and outlook for artificial intelligence in aviation field design application.