Transfer Learning for Aerodynamics Applications
Transfer Learning, Deep Learning, Wind Tunnel Tests
Predicting Onflow Parameters for Domain and Task Adaptation
- A convolutional neural network (ConvNet) model is used to predict angle of attack and onflow speed based on sparse surface pressure data.
- A transfer learning framework utilizing the ConvNet model is proposed for domain and task adaptation.
- Demonstration cases including adaptation to changing data distribution, domain extension, noisy data, and task changes are investigated.
- Published a journal paper: Aerospace Science and Technology (Yilmaz & Bekemeyer, 2026).
Predicting Operating Parameters During Wind Tunnel Tests
- With the goal of bringing the knowledge learned from computational fluid dynamics (CFD) data to the wind tunnel testing phase with streaming data, a transfer learning approach is proposed and demonstrated for a wind turbine blade tip.
- This approach requires first training a neural network offline by using only CFD data, then freezing the weights of the selected layers, and retraining the remaining weights using only wind tunnel data.
- The feasibility of knowledge transfer from CFD runs to wind tunnel tests as well as real-time prediction and online learning during the experiments are successfully demonstrated.
- Published a conference paper: AIAA Aviation’25 (Yilmaz & Bekemeyer, 2025).
