Neural network-based aeroelastic system identification for predicting flutter of high flexibility wings.

Area of Science:

• Aerodynamics
• Aircraft Design
• Aeroelasticity

Background:

• Flutter is a critical and unpredictable instability phenomenon in aircraft.
• Modeling aeroelasticity in high-flexibility wings is complex.
• Existing methods struggle with the unpredictability of flutter.

Purpose of the Study:

• To propose a novel neural network (NN)-based method for aeroelastic system identification.
• To simplify the modeling of high-flexibility wings' aeroelasticity.
• To accurately predict flutter velocity in flexible wings.

Main Methods:

• Developed a neural network (NN) framework for flutter modeling.
• Constructed NN models for high-flexibility wings with varying materials and sizes.
• Utilized system identification techniques for flutter velocity prediction.

Main Results:

• The NN-based method effectively predicts flutter velocity for flexible wings.
• The approach simplifies the complex modeling of aeroelasticity.
• Accuracy was validated against simulation results.

Conclusions:

• The proposed NN-based system identification is a viable method for predicting flutter velocity.
• This technique offers a simplified approach to modeling complex aeroelastic behaviors.
• The method enhances the potential for safer aircraft design.