Nonlinear control-oriented modeling for very flexible aircraft

A low-order model for control design of very flexible aircraft (VFA) that accounts for geometrical nonlinearities under large deformations is derived in this work. The low-order model uses strain-based modal amplitudes as the flexible states and body-fixed frame linear and angular velocities for the basic flight dynamics. The model is informed by a set of nonlinear functions with explicit physical interpretations, representing the nonlinear structural and aerodynamics properties of VFA. These nonlinear functions can be obtained from higher fidelity structural and aerodynamic models or experimental tests. In this work, these nonlinear functions are captured by a set of neural networks, which are trained based on the data generated using nonlinear static results from the University of Michigan’s Nonlinear Aeroelastic Simulation Toolbox (UM/NAST) solver. The resulting nonlinear VFA model is verified against UM/NAST dynamic simulation results using a representative high-aspect-ratio-wing transport aircraft model.