Predictive Control for Alleviation of Gust Loads on Very Flexible Aircraft

Abstract

In this work the dynamics of very flexible aircraft are described by a set of nonlinear, multi-disciplinary equations of motion. Primary structural components are represented by a geometrically-exact composite beam model which captures the large dynamic deformations of the aircraft and the interaction between rigid-body and elastic degrees-of-freedom. In addition, an implementation of the unsteady vortex-lattice method capable of handling arbitrary kinematics is used to capture the unsteady, three-dimensional flow-field around the aircraft as it deforms. Linearization of this coupled nonlinear description, which can in general be about a nonlinear reference state, is performed to yield relatively high-order linear time-invariant state-space models. Subsequent reduction of these models using standard balanced truncation results in low-order models suitable for the synthesis of online, optimization-based control schemes that incorporate actuator constraints. Predictive controllers are synthesized using these reduced-order models and applied to nonlinear simulations of the plant dynamics where they are shown to be superior to equivalent optimal linear controllers (LQR) for problems in which constraints are active.