Transonic buffet is an aerodynamic phenomenon of self-sustained shock oscillations. The aeroelastic problem caused by it is very complex, including two different dynamic modes: forced vibration and frequency lock-in. The vibration of the structure has a negative influence on the fatigue life of the aircraft. Especially in the region of frequency lock-in, the limit cycle oscillations occur due to the instability of the structural mode. Researchers have accurately predicted the region of frequency lock-in in transonic buffet and have clarified its mechanism by using a linear aerodynamic model. However, the nonlinear aeroelastic modeling and prediction of the transonic buffet remain to be solved. The long short-term memory (LSTM) deep neural network is suitable for predicting the time-delayed effects of unsteady aerodynamics. And it has achieved remarkable results in sequential data modeling. In the present work, a nonlinear model is developed for the aeroelastic system with NACA0012 airfoil in transonic buffeting flow and validated with the coupled computational fluid dynamics/computational structural dynamics (CFD/CSD) simulation. First, the data set and the loss function are specially designed. Then, the reduced-order model (ROM) based on the LSTM of the flow is built by using unsteady Reynolds-averaged Navier–Stokes computations data in a post-buffet state. By coupling the ROM and the single degree-of-freedom equation for the pitching angle, the nonlinear aeroelastic model is finally produced. The results show that the phenomenon of frequency lock-in and the self-sustained buffeting aerodynamics are precisely reconstructed. And the model has a strong generalization ability and can reproduce complex vibrations caused by competition between different modes. In short, the model can replace the CFD/CSD method in the current case with high efficiency and accuracy. The method can be used for modeling and prediction of other various complex aeroelastic systems.
Nonlinear Aeroelastic Prediction in Transonic Buffeting Flow by Deep Neural Network
AIAA Journal ; 61 , 6 ; 2412-2429
2023-06-01
Article (Journal)
Electronic Resource
English
Airfoil Unsteady Aeroelastic Responses in Buffeting Transonic Flows
British Library Conference Proceedings | 2009
|Aeroelastic Responses of Spring-suspended Airfoil Systems in Transonic Buffeting Flows
British Library Conference Proceedings | 2012
|