Search

This paper examines the issues involved in controlling an air-breathing hypersonic vehicle (characterized by their unstable, non-minimum phase dynamics) in the presence of significant modeling uncertainty and nonlinearities (control saturations). Modeling of the vehicle exhaust (plume) is complicated, often requiring computational fluid dynamic (CFD) simulations to capture all relevant effects. The focus of this paper is on obtaining a control law that maintains the vehicle trajectory within an acceptable tube while enforcing control constraints in the presence of modeling uncertainty. A robust domain of attraction based approach is used to generate/validate a feasible tube. The computational aspects of such an approach is examined, and the benefits of a decentralized control technique is considered. This approach is compared with other techniques such as linear matrix inequalities based controller design. These approaches are applied to a command following scenario in order to illustrate the performance of the proposed approach.

In this paper, we examine the control of a scramjet-powered hypersonic vehicle with significant aero-elastic-propulsion interactions. Such vehicles are characterized by open loop unstable non-minimum phase dynamics, low frequency aero-elastic modes, significant coupling, and hard constraints (e.g. control surface deflection limits, thrust margin). Within this paper, attention is placed on maintaining acceptable closed loop performance (i.e. tracking of speed and flight path angle commands) while satisfying hard control surface deflection constraints as well as stoichiometrically normalized fuel-equivalency-ratio (FER) margin constraints. Control surface constraints are a consequence of maximum permissible aerodynamic loading. FER margin constraints are a consequence of thermal choking (i.e. unity combustor exit Mach number) and the fact that thrust loss may not be captured for FER greater than unity. Such limits are particularly important since the vehicle is open loop unstable and “saturation” can result in instability. To address these issues, one can design conservative (i.e. less aggressive or lower bandwidth) controllers that maintain operation below saturation levels for anticipated reference commands (and disturbances). Doing so, however, unnecessarily sacrifices performance - particularly when small reference commands are issued. Within this paper, the above issues are addressed using generalized predictive control (GPC). A 3DOF longitudinal model for a generic hypersonic vehicle, which includes aero-elastic-propulsion interactions, is used to illustrate the ideas.

In this paper, we provide an overview of scramjet-powered hypersonic vehicle modeling and control challenges. Such vehicles are characterized by unstable non-minimum phase dynamics with significant coupling and low thrust or FER (normalized fuel equivalency ratio) margins. Recent trends in hypersonic vehicle research is summarized. To illustrate control system design issues and tradeoffs, a generic nonlinear 3DOF longitudinal dynamics model capturing aero-elastic-propulsive interactions for wedge-shaped vehicle is used. The model is analyzed over a broad range of hypersonic flight conditions (i.e. operating points). The paper highlights how vehicle level-flight static (trim) and dynamic properties change over the trimmable air-breathing corridor (∼ Mach 4.75-12.6, 70-115 kft). Particular attention is paid to thermal choking constraints imposed on control system design as a direct consequence of having a finite FER margin. The dependence of FER margin on altitude, Mach, and the bow flow turning angle is discussed. The latter depends on Mach, altitude, angle-of-attack (AOA), and forebody flexing. It is (briefly) discussed how FER margin can be estimated on the basis of Mach, altitude, and AOA if a flexing upper bound is assumed. The implication of this state-dependent nonlinear FER margin constraint as well as that of the right half plane (RHP) zero, associated with the elevator-flight path angle (FPA) map, on control system bandwidth (BW) and FPA tracking are discussed. It is argued that while the non-minimum phase zero limits the achievable closed loop FPA BW, FER coupling into FPA can be used to address this. This, however, is limited by FER margin limits and may impose constraints on the size of the FPA (and velocity) commands that can be followed. This is particularly important because the vehicle is inherently unstable which implies a closed loop system (with a finite downward gain margin) that can become destabilized if driven sufficiently deep into control saturation. A consequence of this is that designers must take note of the fact that FPA commands which are sufficiently large and/or rapid may be impossible to follow with the desired level of fidelity. This is quantified within the paper. Speed command following issues are also discussed.