Development of a Machine Learning-based Rapid Prediction Model for Aircraft Aerodynamic Characteristics

Abstract

In response to the issues of long computational time and insufficient accuracy in traditional aerodynamic characteristic prediction methods, this study proposes a rapid prediction model based on deep neural networks. Through reasonable network structure design and optimization of key parameters, the model achieves high-accuracy predictions for the aircraft's lift coefficient, drag coefficient, and pitching moment coefficient. The model demonstrates excellent predictive performance under both standard and boundary conditions, exhibiting strong generalization ability and computational efficiency. Test results show that the model significantly improves prediction speed while maintaining high accuracy compared to traditional methods, providing a practical solution for the rapid prediction of aircraft aerodynamic characteristics.