Abstract
Computer Fluid Dynamics (CFD) has become an important method of aircraft design. With the improvement of aircraft performance, the contradiction between accuracy and efficiency of numerical simulation becomes more and more obvious. In this paper, a deep learning framework based on large langrage model Bert” is proposed for predicting pressure and velocity distributions for 3D configurations. Based on our framework, the pressure and velocity distribution can be obtained quickly by inputting the aircraft shape points and incoming flow conditions. The proposed framework avoids the influence of computational grid on neural network model, and can use any CFD or experimental results for training. The transonic states of 500 shapes are calculated by Euler equation as inputs of neural network, of which 400 are used for training and 100 are used for testing. The results show that our method can accurately predict the surface pressure and velocity distribution of aircraft, and the time consumption is only seconds, thus achieving a win-win situation of accuracy and efficiency.
Original language | English |
---|---|
Journal | ICAS Proceedings |
State | Published - 2024 |
Event | 34th Congress of the International Council of the Aeronautical Sciences, ICAS 2024 - Florence, Italy Duration: 9 Sep 2024 → 13 Sep 2024 |
Keywords
- aerodynamics
- deep learning
- large langrage model
- pressure prediction