Unsteady Reynolds-Averaged Navier-Stokes-Based Hybrid Methodologies for Rotor-Fuselage Interaction

Abstract

L IFTING bodies produce wakes that interact with other bodies immersed in the same fluid. In particular for rotorcraft, the problem becomes significantly more complicated since the rotor wake remains near the vehicle in hover, descent, and low-speed forward flight. The proximity of the wake alters the inflow distribution at the rotor andmodifies the helicopter thrust.Moreover, since the main rotor wake may impinge on the fuselage, such interactions are an important consideration in modern rotorcraft design. For example, empennage impingement may result in undesirable handling qualities such as low-speed pitchup and tail buffet. Moreover, the wake can also generate unsteady impulsive loads on the fuselage, resulting in vibrations, thus negatively impacting the crew and passenger flight experience. Given the complexity of rotorcraft interactional aerodynamics problems, it is common for tail and empennage designs to be modified significantly after first flight [1]. Development of many aerospace technologies, not limited to helicopter rotor–fuselage applications, requires accurate resolution of both nearand far-field flow phenomena. Numerical prediction of wakes involve a tradeoff between accuracy, turnaround time, and computational expense [2]. Current grid-based computational fluid dynamics (CFD) codes can theoretically model the entire flowfield, but resolution and preservation of wake features become difficult since typical grid sizes used in industrial simulations are susceptible to numerical dissipation. The artificial diffusion of vorticity that results can be mitigated using grid adaptation techniques and higherorder methods [2–4], but this may not be practical for all applications since computational cost increases significantly. For this reason, computationally efficient hybrid methods may be more attractive, especially during design and for flight-test support. Traditional Lagrangian free-wake methods are inexpensive but become less accurate when vortex elements in the wake become distorted and tangled due to interactions with other vortices and solid bodies (i.e., rotor blades and the fuselage) [5]. These interactions typically occur in the rotor near field, which motivates coupling to a CFD solver to resolve the highly viscous and possibly compressible flow near the rotor. In such an approach, the CFD code does not have to resolve the entire wake region; thus, the size of the CFD domain can be greatly reduced and computational efficiency maximized. Additional challenges associated with surface interactions arise when modeling problems such as rotor–fuselage interactions. The ability of two approaches to hybridize a Reynolds-averaged Navier–Stokes (RANS) CFD solver and a free-wake method for the rotor–fuselage interaction problem is investigated. Predictions are compared with experimental data, as well as prior numerical predictions made with individual code simulations.