1 of 1

Objective: To determine whether aerodynamic loads on the BIRE stabilators during flight maneuvers can be harnessed to recover mechanical energy

Neural Network Models

Energy Study

Simulate flight cases to extract torques and deflections

Train neural networks on flight case dataset to predict intermediate cases

Compute instantaneous energy usage/regeneration from stabilator motion

Apply sortie database to estimate lifetime energy usage

BIRE & NAELL

BIRE

• Vertical stabilizer is removed on fighter aircraft and replaced with rotating empennage
• Reduces weight and drag

​

NAELL

• Numerical Aeroelastic Lifting Line solver used to predict forces and moments on fixed wing aircraft
• Approximately 1,500 simulated flight cases

​

​

​

​

• Each simulation outputs stabilator deflection and hinge torque

Steady Level Flight

Pull-up

Push-over

Push-over

Coordinated turn

Neural Network Structure

• Three input, dual output multi-layer perceptron regressor model
• Inputs:
• Altitude
• Mach Number
• Load Factor/Roll Rate/Bank Angle

​

Pull-up Actual Torque vs Neural Network Predicted Torque

• Outputs:
• Hinge Torque
• stabilator Deflection

Shaft

speed