Effect of Tip Geometry on a Hovering Rotor in Ground ...· Effect of Tip Geometry on a Hovering...

download Effect of Tip Geometry on a Hovering Rotor in Ground ...· Effect of Tip Geometry on a Hovering Rotor

of 18

  • date post

    27-Jun-2018
  • Category

    Documents

  • view

    214
  • download

    0

Embed Size (px)

Transcript of Effect of Tip Geometry on a Hovering Rotor in Ground ...· Effect of Tip Geometry on a Hovering...

  • Effect of Tip Geometry on a Hovering Rotor in

    Ground Effect:

    A Computational Study

    Tarandeep S. Kalra,

    Vinod K. Lakshminarayan,

    James D. Baeder

    A compressible Reynolds Averaged Navier Stokes (RANS) is used to simulate a hov-ering rotor with four different tip shapes (rectangular, swept, BERP-like and slotted) tounderstand the effect of rotor wake in-ground effect(IGE). The simulations are built uponan original framework developed by the current authors. The tip vortices are better re-solved by modifying grids in order to follow the path of the rotor wake and using oversetmeshes to better preserve the vortices. Overall the fully turbulent RANS computationspredict a highly diffused vortex core compared to experiments for the rectangular, sweptand BERP-like tip. The eddy viscosity levels are quite high and it is believed that thecurrent formulation of turbulence modeling is leading to excessive diffusion of vorticity. Tocircumvent this problem, simulations are performed with the laminar flow assumption onthe wake capturing meshes. The tip vortices are captured well by using this assumptionand the tip vortex characteristics agree well with the experiments. Between the differenttip shapes, the rectangular and swept tip show similar vortex strengths. The BERP-liketip has a slightly diffused vortex core than the former two tip shapes. The slotted tip showsa highly diffused vortex core compared to the other three tips.

    Nomenclature

    A Area of the rotor blades (R2)AR Aspect Ratio, AR = R/cc Rotor blade chordCP Rotor power coefficient in-ground effect = Power/(R

    2V 3tip)CT Rotor thrust coefficient in-ground effect = Thrust/(R

    2V 2tip)Mtip Tip Mach numberr Radial distancerc Vortex core radiusR Radius of the rotorRe Reynolds numberVtip Tip speedV Swirl velocityz Rotor height above ground Circulation Wake age (degrees)o Collective pitch (degrees)

    Graduate Research Assistant, tkalra@umd.edu Postdoctoral Fellow, vinodkl@stanford.edu Associate Professor, baeder@umd.edu

    1 of 18

    American Institute of Aeronautics and Astronautics

    Dow

    nloa

    ded

    by S

    TA

    NFO

    RD

    UN

    IVE

    RSI

    TY

    on

    July

    19,

    201

    3 | h

    ttp://

    arc.

    aiaa

    .org

    | D

    OI:

    10.

    2514

    /6.2

    013-

    2542

    31st AIAA Applied Aerodynamics Conference

    June 24-27, 2013, San Diego, CA

    AIAA 2013-2542

    Copyright 2013 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

  • I. Introduction

    The brownout phenomenon consists of the creation of a dense dust cloud that engulfs the rotorcraft duringin-ground effect operation. Apart from the obvious problem of rendering the pilot visually disoriented, theseclouds can also be responsible for blade erosion and mechanical wear. Current day workarounds to thebrownout problem include the use of sensors and avionic displays for improving the situational awareness orthe employment of piloting strategies to avoid the brownout cloud. Although these solutions can contributein limiting the number of brownout related mishaps, these strategies only provide a temporary solution anda more permanent solution is desired for this problem. Since the interaction of the rotor wake with thedust particles on the ground is the driving force of the brownout phenomenon, it is believed that a goodunderstanding of the underlying flow physics can provide insight to develop effective means of preventingand/or mitigating the adverse effects of rotorcraft brownout.

    Over the last few years, significant strides have been made in the experimental analysis of the brownoutproblem in controlled environments. Detailed particle image velocimetry (PIV) measurements of the flowfieldperformed by Lee et al.1 at micro scale and Milluzzo et al.2 at subscale showed a whole host of aerodynamicphenomena such as diffusion, vortex stretching and turbulence generation occurring close to the ground.PIV measurements enabled the experimentalists to measure the velocity profiles close to the ground for aquantitative estimate of the flowfield. Recent experiments have been performed by Hance et al.3 to includethe effect of fuselage on the wake of a two-bladed subscale rotor. Additional experiments conducted byJohnson et al.4 and Sydney et al.5 showed the phenomenon of entrainment and uplift of sediment particlesin the presence of the rotor wake of a hovering micro scale rotor. Other experimental studies include thePIV measurements and brownout simulations performed by Nathan and Green.6 These experiments wereconducted in a wind tunnel to simulate low speed forward flight on a micro scale rotor.

    These aforementioned experiments provide a detailed understanding of the fluid dynamics for micro andsubscale rotor in-ground effect. However, the wide range of parameters affecting the brownout problem arestill difficult to model in experiments. The main experimental difficulty comes in simulating the brownoutconditions in full-scale rotors. Computational studies can be used to overcome these challenges by simulatinga wide range of brownout conditions with relative ease. However, computational studies have their own shareof challenges such as:

    The tip vortex emanating from the blade needs to be preserved for a significantly long time to capturetheir interaction with the ground, and therefore, can be computationally expensive, especially whenthe ground distances are large.

    It is necessary to resolve the boundary layer and the turbulence at the ground, which can be extremelychallenging especially using non-CFD computational methods.

    Modeling the fluid-particle interaction for these complicated flow-field is extremely challenging. Devel-opment of a particle model is still an active field of research.

    These challenges are further aggravated in terms of computational expense when the scale of the problemis increased. To simplify the computations, most recent studies simplify the multi-phase brownout problem byignoring the effects of the airborne sediment particles on the flowfield and simulate only the effect of flowfieldon the particles. This is a reasonable approximation for the brownout problem, except possibly very close tothe ground where the particle density cloud may be large enough to affect the momentum of the fluid. Thisdecoupled approach allows one to concentrate on the problem of tracking the evolution of the rotor tip vorticesand its interaction with the ground boundary layer. Over the last few years, researchers modeling brownouthave employed aerodynamic models of various levels of sophistication. Most approaches track the tip vortexin the particle-fixed Lagrangian frame of reference.79 Even though these free-vortex type methods can trackthe vortex for a long distance, these methods involve a certain degree of empiricism in determining the vortexcore radius and roll-up. Additionally, these methods rely on approximate sublayer models to predict theground boundary layer and thus may be best suited to provide qualitative prediction of the ground boundarylayer. Methods using the vorticity transport model (VTM)1012 or vorticity confinement13,14 provide acapability for a more fundamental representation of the formation and evolution of the tip vortex. However,VTM applications to brownout have been run as an inviscid model and thus still require an approximationto the sublayer to capture the ground boundary layer. On the other hand, vorticity confinement applicationsto brownout have used very coarse meshes near the ground that may not be sufficient to capture the details

    2 of 18

    American Institute of Aeronautics and Astronautics

    Dow

    nloa

    ded

    by S

    TA

    NFO

    RD

    UN

    IVE

    RSI

    TY

    on

    July

    19,

    201

    3 | h

    ttp://

    arc.

    aiaa

    .org

    | D

    OI:

    10.

    2514

    /6.2

    013-

    2542

    Copyright 2013 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

  • of the wall jet and interaction with the vortical wake. Furthermore, in these studies the blades were treatedusing lifting-line aerodynamics with their accompanying assumptions. In another work, Thomas et al.15

    presented a hybrid methodology combining the capabilities of a high-fidelity RANS solver with a free-wakemethod to simulate the single-phase and two-phase flowfield environment beneath a hovering rotor. Thismethod requires lesser computational expense compared to a full RANS single-phase simulation but stillrelies on empirical correlations in the free-wake domain of the flowfield. Thus, all the above mentionedapproaches in their current forms still retain a fair amount of empiricism.

    A RANS-based CFD solver with sufficient mesh resolution can resolve the detailed velocity profiles nearthe ground without using empirical vortex core factors or a sublayer model. The primary disadvantage of thismethodology is the computational expense required to capture the evolution of the wake structure. However,with parallelization and the use of intelligent clustering of mesh points in regions of interest, one can usethese high fidelity methods to simulate the brownout problem. Earlier work by current authors 16 and 17

    showed the capability of this methodology to provide accurate performance and flowfield characteristics ofthe micro scale rotor tested by Lee et al.1 The current work extends this methodology to simulate the one-bladed subscale rotor experiments of Milluzzo et al.2 The simulations investigate the effect of four differentblade tip shapes on the tip vortex formation, evolution and its interaction with the ground.

    II. Experimental Setup for Validation

    The experimental setu