Adaptive, neural and robust control of wing-rock and aeroelastic system

Modern aircraft exhibit wing-rock phenomenon and aeroelastic instability. Wingrock (roll single degree of freedom motion) and aeroelastic systems' (two degrees of freedom) behavior are described by complex nonlinear differential equations. The nonlinearities in the dynamics of these systems give rise to limit cycle oscillations beyond critical speed of aircraft. The onset of wing-rock and aeroelastic instability limits the performance of aircraft and can even lead to catastrophic consequences. Therefore, control of wing-rock motion and stabilization of aeroelastic systems are important. In the past, several studies have been made and experimental and analytical results have been obtained to explain the wing-rock and aeroelastic phenomena in wind-tunnel tests, and also control systems have been derived.;Motivation for this research is the importance of flying aircraft in a large flight envelope in which complex uncertain aerodynamic nonlinearities appear, causing instabilities and flutter in the aircraft wings. For the control of wing-rock motion and the stabilization of aeroelastic instabilities, new control systems are designed. Because modeling of nonlinear dynamics of wing-rock motion and aeroelastic systems are imprecise, the control algorithms must be insensitive to model uncertainties. Apparently control theory for deterministic systems is not applicable to uncertain systems.;For the stabilization of wing-rock, two non-certainity equivalent adaptive (NCEA) laws are designed. The first control system includes a finite form realization of a speed-gradient adaptation law, and the second controller is based on the Immersion and Invariance (I&I) theory. For the nonlinear multi-input multi-output (MIMO) aeroelastic systems, equipped with leading- and trailing-edge control surfaces, four distinct control systems are designed. First, a Chebyshev neural adaptive control law is derived for the suppression of limit cycle oscillations (LCOs) of the prototypical wing. For this derivation SDU decomposition of the high-frequency constant gain matrix is utilized for obtaining a singularity free controller. Then for a multi-input aeroelastic system with state dependent input matrix, a higher-order robust sliding mode control law for finite-time stabilization is derived. This is followed by the design of a suboptimal controller based on the state-dependent Riccati equation (SDRE) method. Finally, a suboptimal control law is designed for the control of the aerolelastic system, based dierential game theory. In this approach, the wind gust is treated as an adversary which tries to destabilize system. These control algorithms are simulated using MATLAB and SIMULINK to verify their performance. Results show that the designed controllers are effective in suppressing the limit cycle oscillations.