AIAA 2010-4238
28th AIAA Applied Aerodynamics Conference
28 June - 1 July 2010, Chicago, Illinois
Rotor Configurational Effect on Rotorcraft Brownout
Sayan Ghosh,∗ Mark W. Lohry∗ and R. Ganesh Rajagopalan†
Iowa State University, Ames, Iowa, 50011, USA
One of the major problems faced by rotorcraft undergoing aerodynamic maneuvers at
low altitude in arid desert regions is in-flight visibility restriction due to the formation
of dust clouds in the air. This phenomenon known as Brownout is caused by dust and
sand particle entrainment into the air by the impinging downwash of the rotor. One of
the various factors which influence the characterization of brownout is rotor configuration.
This paper describes the analysis of rotor configurational (single-rotor, tandem-rotor, tiltrotor, and quad-rotor) effect on ground footprint, wake and brownout using Computational
Fluid Dynamics. In this study, the flow field is calculated by solving Reynolds averaged
Navier-Stokes equation using the SIMPLER algorithm. To solve for turbulence properties,
Realizable k − equation is used. The rotor is modeled as momentum source whereas the
dust behavior is solved using an Eulerian-based dust transport equation.
Nomenclature
ui
Fluid velocity, m/s
upi Particle velocity, m/s
dp
Particle diameter, m
ρ
Fluid density, kg/m3
ρd
Dust density, kg/m3
p
Fluid pressure, N/m2
g
Acceleration due to gravity, m/s2
µ
Fluid molecular viscosity, kg/(s · m)
µt
Turbulent eddy viscosity, kg/(s · m)
µef f Effective viscosity, kg/(s · m)
µtp
Turbulent particle viscosity, kg/(s · m)
Γd
Particle dispersion coefficient, m2 /s
Re Fluid Reynolds number
Rep Particle Reynolds number
k
Turbulent kinetic energy, m2 /s2
Turbulent dissipation, m2 /s3
i
Srotor Rotor source term
Sd
Dust source term
R
Rotor radius, m
c
Rotor blade chord
Nb Number of blade in rotor
wt
Particle terminal velocity, m/s
u∗
Fluid friction velocity, m/s
u∗t Particle threshold velocity, m/s
CD Drag coefficient
CL Lift Coefficient
CT Coefficient of thrust
∗ Graduate
† Professor,
Student, Department of Aerospace Engineering, Iowa State University.
Department of Aerospace Engineering, Iowa State University, and AIAA Associate Fellow.
1 of 19
American
of and
Aeronautics
and
Astronautics
Copyright © 2010 by Sayan Ghosh. Published by the American
Institute ofInstitute
Aeronautics
Astronautics,
Inc.,
with permission.
I.
Introduction
otorcraft brownout has become one of the major sources of accidents while operating in desert-like
R
regions. The particle entrainment into the air causes reduction in visibility and situational awareness
for the pilot, as well as damage of engine components and rotor blades. The entrainment of particles and
formation of dust clouds starts at low altitude flight but the situation worsens during take-off and landing.
At present, there is lack of complete knowledge of physics involving aerodynamics and particle entrainment
but studies are being made to analyze the effect of various parameters on the brownout phenomenon. One
such parameter is rotor configuration. Various rotor configurations have different in-ground effect which is
caused by complex aerodynamic interaction of wakes from different rotors and depends upon their relative
orientation. Each configuration has different wake patterns and ground signatures which directly affect the
entrainment process of particles. Though complete mitigation of brownout might not be possible, proper
understanding of the physics and the effect of rotor configuration might help in controlling the extent of dust
cloud formation.
Some flight test data on dust distribution for different aircraft are available in Rogers1 and Chatten.2
However, in order to computationally simulate this phenomenon, it is first imperative to correctly predict
the rotor downwash in different flight conditions. Also important is understanding the physics of particle
entrainment and distribution in the brownout dust cloud. Some knowledge can be obtained from the field
of riverine and aeoline sedimentology. It has been found that the empirical models used in describing the
behavior of suspended particles in water or air can also be applied to the helicopter brownout problem.
Two approaches are popular in studying particle entrainment and transport processes. The first is the
Lagrangian approach3 in which the trajectories of individual particles are tracked and these particles are
taken to represent a dust cloud. The dynamics of each individual particle is modeled. Thus, for the overall
flow to be reliably represented, a large number of particles must be modeled, thereby making this approach
computationally very expensive. In the second approach, the so-called Eulerian approach, overall dynamics
of the particle distribution in the air is modeled using suitable transport equations. This approach has been
applied to the brownout problem by Ryerson,4 and Haehnel.5 Both these works use a two-phase model to
represent the particle dynamics and assume a one-way coupling between the fluid and the particles, i.e.,
the fluid affects the particle but not the other way around. Another documented study of the brownout
problem was conducted by Phillips and Brown.6 This work used the Vorticity Transport Model (VTM) to
solve the rotor flowfield and coupled it with an Eulerian-based dust Transport equation to represent the dust
entrainment process. Brown’s dust transport model uses empirical correlations borrowed from sedimentology
and studies the evolution of particle density distribution and its relation with flowfield velocity and vorticity.
This study found the interaction of the tip vortices with the ground is by far the most important contributor
to the formation of the initial dust cloud. Therefore, the vorticity and velocity profile in the rotor wake are
crucial and Phillips et al.7 determined these are affected by parameters like solidity and blade twist. These
suggest the geometric properties of the rotor have a great influence on the type and extent of the dust cloud
formed.
Another approach,8, 9 employs a real-time free wake model, coupled with a dust and debris model, that
uses analytical methods in a Lagrangian framework. This model results in a real-time brownout simulator
that gives pilots a feel for the brownout cloud formed under different flight conditions, ground cover and
type of aircraft. A visual obscuration model, based on light a scattering method, is also used to visualize
brownout and train pilots accordingly. However, the Lagrangian approach makes this approach computationally expensive.
One of the important factors in developing a dust transport model that can be used in conjuction with
a Computational Fluid Dynamics (CFD) code is the development of a suitable “entrainment function” to
describe the entrainment flux. Haehnel10 studied the physics of particle entrainment under the influence of
an impinging jet and determined the driving force for particle entrainment is the Reynold’s stress associated
with the turbulent fluctuations of the flow. Since an impinging jet is similar to the wake of the rotor in
ground effect, it can be deduced that solving correctly for the turbulent fluctuations in the rotor’s flowfield
is an important step in solving the brownout problem. Therefore, there is scope for developing a complete
brownout model that incorporates all the components of brownout as observed in reality.
In this paper, an attempt has been made to study the physics of brownout at low altitude hovering flight,
with the focus on the effect of rotor configuration using computational tools. For the study, full representative
rotorcraft configurations are used including mock-up fuselage geometry. Since the rotor flowfield is highly
turbulent, a RANS based flow solver has been used to solve for the flow field; wherein, the rotor is modeled
2 of 19
American Institute of Aeronautics and Astronautics
as rotor momentum source,11 thus, making the solver efficient, simple and robust. The dust transport during
the brownout is modeled using Eulerian-based dust transport model. The entertainment of particles near
the ground is modeled empirically, using mathematical relations from geophysical research.
II.
Governing Equations and Methodology
During rotorcraft brownout, multiple length scales dictate the physics of the simulation, which involves
aerodynamic processes like ground effect on rotor wake, turbulent burst, surface wall jet, tip vortices, and
their interactions. These aerodynamic interactions, coupled with the interaction of the particles, become
a complex multi-phase phenomenon. Accuracy of the simulation, particularly close to the ground where
particle entrainment takes place, is highly dependent upon the turbulence characteristics and ground shear
predicted by the mathematical model used.
To accurately and economically predict the turbulent flow field, an incompressible Reynolds Averaged
Navier Stokes (RANS)-based solver has been used in the Cartesian coordinate system. The equations for
conservation of mass and momentum of turbulent incompressible rotor flow field are given by
∂ui
=0
∂xi
∂ρui uj
2 ∂k
∂p
∂
∂ui
∂ρui
i
+
µef f
−
=−
+
+ Srotor
∂t
∂xj
∂xi
∂xj
∂xj
3 ∂xi
(1)
(2)
where ρ is density, xi is position, ui is mean velocity, p is pressure, k is turbulent kinetic energy, and t is
time. µef f is effective viscosity which is given by sum of laminar viscosity, µ, and turbulent viscosity, µt , as
i
is the momentum source used to model the rotors.
µef f = µ + µt . The term Srotor
Turbulent viscosity depends upon flow properties and is a function of turbulent velocity and length scale.
To solve for the turbulent properties, RANS-based Realizable k − model12 is used to solve for turbulent
kinetic energy, k, and turbulent dissipation, . From these turbulent properties, the turbulence length scale,
lt , is calculated as k 3/2 / and velocity scale, Vt , is approximated as k 1/2 . The turbulent viscosity can be
calculated as
k2
µt = f (lt , Vt ) = Cµ ρ
(3)
where Cµ is a factor relating turbulent viscosity with length scale, and velocity scale and is given by
Cµ =
1
∗
A0 + As kU
(4)
q
where U ∗ ≡ Sij Sij + Ω̃ij Ω̃ij , Ω̃ij = Ωij − 2ijk ωk , Ωij = Ωij − ijk ωk . Ωij is the mean rate-of-rotation
tensor viewed in a rotating reference frame with
√
√ the angular velocity ωk .
The constants A0 and As are 4.04 and 6 cosφ, respectively, where φ = 13 cos−1 ( 6W ), W =
p
Sij Sjk Ski
∂u
∂ui
, S̃ = Sij Sij , and Sij = 12 ∂xji + ∂x
.
j
S̃ 3
Transport equation for turbulent kinetic energy is given by
∂
∂
∂
µt ∂k
(ρk) +
(ρkui ) =
µ+
+ Pk − ρ
(5)
∂t
∂xi
∂xi
σk ∂xi
where Pk is the production term for turbulent kinetic energy and ρ is the destruction term due to turbulent
dissipation.
The production term Pk can be related to the mean rate of strain tensor, S, and vorticity, Ω,13 as
Pk = µt SΩ
The equation for turbulent dissipation is given by
∂
∂
µt ∂
2
∂
√
(ρ) +
(ρui ) =
µ+
+ ρ C1 S − ρ C2
∂t
∂xi
∂xi
σ ∂xi
k + ν
h
i
η
where C1 = max 0.43, η+5
and η = S k .
The values of the coefficient used in the Realizable k − model is given in Table 1.
3 of 19
American Institute of Aeronautics and Astronautics
(6)
(7)
Table 1. Closure coefficient for Realizable k − model
C1
C2
σk
σ
A.
1.44
1.9
1.0
1.2
Rotor Source Model
One of the major factors influencing the analysis is accurate prediction of rotor wake. There are a number
of sophisticated computational algorithms to model the rotor flow which can predict all the possible regimes
of flow and aero-elastic behavior of rotors. Since our interest is on the overall flow field rather that the flow
near the rotor blades, a simple, robust, and efficient method developed by Rajagopalan11 is used, in which
i
the rotors are modeled as source terms in the momentum equation. The momentum source term, Srotor
, is
added to the momentum equation at the point where the blade passes. The magnitude of the source term
is a function of rotor geometry, blade cross-section aerodynamic characteristics, and the aerodynamic forces
exerted by the rotor blades.
i
i
~ , Ω, x, y, z, ρ, µ, Re, M, b, c)
Srotor
= Srotor
(Cl , Cd , α, α̇, V
(8)
To calculate the rotor source term, the rotor
is discretized into a span-wise element. The blade
properties, like chord length, airfoil thickness, plane
deflection, and cross-sectional area are assumed to
be constant over the element. The forces acting on
the blade section are shown in figure 1. From the figure, the normal, fn , and tangential, ft , force acting
on the rotor disc plane are given by
fn
= L cos − D sin fθ
= L sin + D sin (9)
Figure 1. Aerodynamics forces at a blade section
where is the inflow angle given as
= arctan(−vn0 /vθ0 )
vn0
(10)
vθ0
where
and
are the transformed normal and the tangential velocities in the rotor disc plane. Lift, L,
and drag, D, are given by
L =
D
=
1 02
ρv Cl c ds
2
1 02
ρv Cd c ds
2
where c is the airfoil chord. v 0 is the relative velocity seen by the airfoil secion given by
q
v 0 = v 0 2n + v 0 2θ
(11)
(12)
Cl is the section coefficient of lift, and Cd the coefficient of drag, which are obtained from an airfoil lookup
table by calculating the local angle of attack, α, and local Mach number, M 0 . Local angle of attack is
calculated as
α=φ−
(13)
where φ is the twist angle at that section which is determined from rotor geometry. Mach number is given
by
M 0 = v 0 /a
(14)
where a is the speed of sound. Determining the aerodynamic coefficients by this manner implicitly allows
the compressibility effect on the aerodynamic characteristics of the blade section.
4 of 19
American Institute of Aeronautics and Astronautics
Once the forces on the airfoil section are calculated, the resultant force can be transformed to the
computational domain to obtain the resultant force acting on the blade F~ . Thus, the force acting on the
fluid, −F~ , is then time averaged and is added to the grid cells in which the rotor is passing as a momentum
i
source term, Srotor
.
Using this method, explicit details of each rotor blade and chordwise flow over rotor blade are not resolved.
The calculation of rotor source is independent of flowfield and does not depend upon characteristics like
incompressibility and turbulence. More details can be found in Rajagopalan and Mathur.14
B.
Dust transport model
During rotorcraft brownout, the aerodynamic interaction of rotor wake with the dust particle can be treated
as a multi-phase problem in which air is the carrier phase and dust is the discrete phase. From the field
studies achieved by Chatten2 on brownout cloud characteristics on six rotorcraft airframes at the Yuma
Proving Ground (YPG), it was found that brownout dust clouds mainly contain particles ranging from 2 to
350µm, with the mass concentration at rotor tip location ranging from 10mg/m3 to 700mg/m3 . With the
assumption that the particle bulk density ratio is very low (less than 1%) for most of the rotor flow field, the
dust transport model can be loosely treated as one-way coupled, i.e. the flow field affects the dust behavior
where as the dust particle does not affect the flow field. The assumption of one-way coupling may not be
accurate in the region near the ground where the dust concentration is high due to the entrainment process,
but since the region of high dust concentration is small compared to the overall flow field, the assumption
of one-way coupling is justifiable.
The dust transport is solved in an Eulerian frame of reference where the dust is assumed to be a continuum. The mass transport equation of dust can be calculated by summing the mass change of the dust in the
control volume to net efflux of mass from the control surface and the rate of dust generation in the volume.
∂
∂
∂
∂ρd
∂
(ρd ) +
(ρd upi ) −
(ρd wt ) = Sd +
Γd
(15)
∂t
∂xi
∂x3
∂xi
∂xi
where ρd is the dust density, up is dust particle velocity, wt is the terminal velocity of dust particles, Γd is
the dispersion coefficient of dust particles due to turbulence, and Sd is dust source term.
Particle terminal velocity is the particle to fluid relative velocity, when the aerodynamic forces balance
∂
the gravitational force on the particle. Thus the third term, ∂x
(ρd wt ), takes care of the settling effect on a
3
particle, due to gravity. Relation of particle terminal velocity is given as
wt =
4dp σp
3CD (Rept )wt
(16)
where σp = ρp /ρ is the particle density ratio. Rept = wt dp /ν is particle Reynolds number at terminal
velocity and CD is particle drag coefficient which is a function of particle Reynolds number,
CD =
24
fdrag
Rep
where fdrag is known as the drag function and can calculated as15

1 + 0.15Re0.687 if Re ≤ 1000
p
p
fdrag =
0.0183Re
if Rep > 1000
(17)
(18)
The particle dispersion coefficient is calculated as16
Γd =
µtp
σpt
(19)
where µtp is turbulent particle viscosity and σpt is turbulent particle Schmidt number, which varies from
0.34 − 0.7. Turbulent particle viscosity µtp is related to turbulent fluid eddy viscosity µt by
µtp =
µt
1 + ττfv
5 of 19
American Institute of Aeronautics and Astronautics
(20)
where τv is particle velocity response time and τf is fluid relaxation time.
The dust source term, Sd , represents the generation of dust near the ground
due to dust particle entrainp
ment. On the ground, particles lift off when the friction velocity, u∗ = τw /ρ, is higher than the particle
threshold friction velocity, u∗t . Particle threshold friction velocity is the minimum friction velocity required
to give the initial motion to the particle on the ground against the retarding gravitational force and cohesive
forces. The particle threshold velocity depends on various factors like particle size, density, fluid density,
particle friction Reynolds number, etc. Lu and Shao17 proposed that under ideal conditions, u∗t can be
described as
s ρp
a2
1
a1
gdp +
(21)
u∗t =
κ
ρ
ρdp
where κ represents a surface roughness factor. From experiments the value of the coefficients a1 and a2 are
found to be approximately 0.0123 and 3 × 10−4 kgs−2 , respectively.
In general, particles with small terminal velocity goes into suspension as dust whereas relatively larger
particles go into a mode called saltation, in which the particles hop along the surface in the wind direction.
The saltating particles gain momentum from the flow and strike the sand bed with high energy resulting in
further emission of particles. This process is one of the major mechanism of dust emission.
Since the entrainment of dust near the ground is a complicated process, the entrainment of dust is
modeled through dust source term, Sd , the magnitude of which depends upon the vertical dust flux in the
saltation layer. The vertical dust flux near the ground depends upon ground friction velocity and many
empirical relations can be found in the domain of geophysical research. Nalpanis18 found that in the parallel
flow over a dusty flat plate, the vertical dust flux, F , in the saltation layer decreases exponentially with
height as
F = Fo e
−λgz
u2
∗
(22)
where Fo is vertical dust flux at the ground, z is the height from the ground, and λ is a parameter which
varies over 50%. According to a scheme proposed by Marticorena and Bergametti,19 the vertical dust flux
at the ground, Fo , depends upon saltation or horizontal particle flux Q as
Fo = Qe13.4f −6.0
(23)
where f is percentage of clay in soil. For the current work, f = 0.1 was used. Similar to vertical particle
flux, saltation flux also depends on friction velocity u∗ . The saltation flux, Q, is calculated using a model
developed by Bagnold20 as

0
if u∗ < u∗t
Q= c ρ (24)
2
 o u3∗ 1 − u∗t
if u∗ ≥ u∗t
g
u2
∗
t
where co = 0.25 + 0.33w
u∗ .
The particle velocity is calculated as
upi = uri + ui
(25)
where uri is particle relative velocity with respect to fluid velocity. For flows with low particle response
time (0.001 < τv < 0.01s), an algebraic slip formulation 21, 22 is used to calculate particle relative velocity.
Assuming that local equilibrium has been reached between the particle and fluid, the relative velocity can
be given by
τv (ρp − ρ)
uri =
ai
(26)
fdrag
ρp
where ai is acceleration given by the mixture velocity (fluid velocity for low particle volume fraction) as
ai = fi − ui
∂uj
∂ui
−
∂xj
∂t
(27)
where fi is the acceleration due to body force. Since no electrostatic, magnetic effect, or other body forces
have been considered and the effect of gravitational force has already been taken into account in the particle
transport equation using the terminal velocity, wt , f = 0 has been used to calculate particle relative velocity.
6 of 19
American Institute of Aeronautics and Astronautics
In turbulent flow, a diffusion term is added to the relative velocity to accommodate the effect of turbulent
fluctuation.
τv (ρp − ρ)
uri =
ai + utri
(28)
fdrag
ρp
where
utri = µtp
C.
1 ∂ρd
ρd ∂xi
(29)
Numerical approach
The flowfield is discretized in a non-body fitted Cartesian grid. In the solver, the governing conservation
equations of fluid mass and momentum are solved using a finite-volume based method known as SIMPLER.23
In this algorithm, the flow field is determined by solving for primitive variables, namely the static pressure
and the velocity vector, directly from the mass and momentum conservation equations. For any general
variable φ, the transport equations can be discretized using finite volume method in the form of
ap φp = Σanb φnb + b
(30)
where φp is the value of a general variable at cell center p, ap and anb are the coefficients that link φp to
the neighboring cell’s φs. The discretized equation is solved using line-by-line method using TriDiagonalMatrix Algorithm (TDMA) and Gauss-Seidel iterative scheme. During the computational process, the RANS
equation is solved first at each time step. While solving the RANS, the rotor source model is invoked at each
iteration to add the rotor source terms. From the converged flow field, the turbulent properties are solved
through the turbulence models. After the calculation of flow field variables, dust concentration is solved
using the dust transport model. The entire algorithm is given in the flow chart shown in figure 2.
Figure 2. Flowchart of Rotorcraft Brownout Model
III.
A.
Results
Turbulent Rotor Model Validation
To validate rotor source modeling, simulation has been accomplished on a rotor similar to experiments
completed by Rabbott.24 A two-bladed rotor has been used with 4.57m diameter. Simulations were carried
out at various collective pitches (Θ). For the current simulation, NACA0012 airfoil was used. Chord length
was 0.153 times the radius and there was zero twist along the blade. The variation of aerodynamic blade
load with radial locations at various collective pitches is plotted in Figure 3. The rotor source model is able
to perfectly capture the trend of the aerodynamic load along the blade radius.
B.
Brownout model validation
Since there have not been many experimental data on brownouts, it is very difficult to validate the brownout
model quantitatively. The present model has been qualitatively validated with wind tunnel experiments
7 of 19
American Institute of Aeronautics and Astronautics
(a) Θ = 3.0o
(b) Θ = 4.5o
(c) Θ = 8.5o
(d) Θ = 9.2o
Figure 3. Aerodynamic blade load comparisons with Rabbott’s experiment at various Θ
conducted by Nathan and Green25 at University of Glasgow. In their experiment, a small model rotor was
used and was placed at one radius above the ground. The wind tunnel speed was set so the thrust-normalized
advanced ratio, µ∗ , is equal to 0.65.
Thrust-normalized advanced ratio is given by
µ∗ = p
µ
(31)
(CT /2)
where µ is rotor advance ratio and CT is coefficient of thrust.
µ = V cos α/(ΩR)
(32)
where α is tilt angle, Ω is angular velocity, and R is the radius
of rotor. CT is given by
CT =
T
ρ(ΩR)2 A
(33)
where T is rotor thrust and A is rotor disc area.
In their experiment, very fine, powder-like particles were Figure 4. Particle transport by rotor downtunnel experiment (Nathan et
used as dust. The motion of the particle transported by rotor wash in wind
al. (2008)25 )
flow was captured by high speed camera (figure 4).
Simulation was performed on similar rotor flow conditions
using the present brownout model. Dust particle density distribution and velocity are shown in figure 5.
It can be seen the present numerical model was able to capture the same extent of the brownout cloud.
Another remarkable feature captured by the present model is the stagnation region of the dust particle on
the ground at about one radius upstream.
8 of 19
American Institute of Aeronautics and Astronautics
(a) Dust particle Density
(b) Particle velocity vector
Figure 5. Dust particle distribution predicted by brownout model
C.
Rotor Configuration Effect on Brownout
Full rotorcraft configurations with single-rotor, tandem-rotor, tilt-rotor, and quad-rotor have been used to
study rotor configuration effect on brownout (figure 6). In this research, study of brownout is done on
hovering at different heights, which also depicts the scenario of vertical landing and take off. The rotor
properties used are given in Table 2. For all the rotors NACA0012 airfoil has been used. The collective pitch
of each rotor is adjusted so that CT /σ for all the rotorcrafts are the same, while hovering at OGE (out of
ground effect). σ is the solidity of the rotor given by σ = Nb c/πR, where Nb is the number of rotor blades
and c is the blade chord.
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 6. Rotor configuration
Table 2. Properties of rotor for multi-rotor configurations
Single-rotor
Tandem-rotor
Tilt-rotor
Quad-rotor
Rotor Radius
Rotor Blades (per rotor)
Twist
σ
CT
12m
9m
6m
6m
5
3
3
3
12o
12o
16o
16o
0.095
0.083
0.105
0.105
0.0071
0.0063
0.0076
0.0077
The simulation was carried out for four rotor heights 1.0, 1.5, 2.0, and 2.5Rref , where Rref is the rotor
radius of a single-rotor helicopter. Corresponding non-dimensional heights (H/R) for the tandem-rotor are
1.33, 2, 2.66, and 3.66. For tilt-rotor and quad-rotor, the non-dimensional heights (H/R) are 2.0, 3.0, 4.0, and
5.0.
Velocity streamlines colored with velocity magnitude, on mid-rotor lateral plane of single-rotor and tiltrotor are plotted in figures 7 and 9, respectively. For tandem-rotor and quad-rotor, velocity streamlines are
plotted on three different lateral planes, i.e. front rotor-mid plane, fuselage-mid plane, and rear mid-rotor
plane, in figures 8 and 10.
9 of 19
American Institute of Aeronautics and Astronautics
In single-rotor, large recirculating regions are observed underneath the fuselage at almost all the heights.
The fuselage plays an important role in the formation of these recirculating regions. As the rotor height
increases, the effect of fuselage decreases, which can be seen in the decrease of strength and size of the
recirculating region. tandem-rotor show the same trend of behavior, where large recirculating regions are
observed underneath the fuselage on rotor mid planes, at rotor heights of H = 1.0, 1.5, and 2.0Rref . At
height of 2.5Rref , the recirculating region has diminished. On the mid-fuselage lateral plane of the tandemrotor, the interference effect of two counter-rotating rotors is observed, where the wake of the rotor is shifted
towards the starboard side. The strength of the wall jet has also been found to be stronger on the starboard
side of tandem-rotor.
For both the tilt-rotor and quad-rotor, strong recirculating zones are observed just below the fixed wings.
Since there is no overlap of rotors, the interference effect is not that significant as compared to tandem-rotor.
Two counter rotating vortices are observed near the ground for tilt-rotor’s case, formed by merging wakes of
two rotors. In quad-rotor, the wake in the lateral mid-fuselage plane is converging as oppose to the case of
tandem-rotor where the wake diverges. This is due to non-overlapping rotors in the case of quad-rotor. At
a height of 2.5Rref , the recirculating region near the ground is negligible for both tilt-rotor and quad-rotor.
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 7. Velocity streamlines in lateral plane for single-rotor hovering at different heights
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 8. Velocity streamlines in lateral plane for tandem-rotor hovering at different heights
10 of 19
American Institute of Aeronautics and Astronautics
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 9. Velocity streamlines in lateral plane for tilt-rotor hovering at different heights
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 10. Velocity streamlines in lateral planes for quad-rotor hovering at different heights
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 11. Velocity streamlines in longitudinal plane for single-rotor hovering at different heights
11 of 19
American Institute of Aeronautics and Astronautics
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 12. Velocity streamlines in longitudinal plane for tandem-rotor hovering at different heights
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 13. Velocity streamlines in longitudinal plane for tilt-rotor hovering at different heights
(a) H = 1.0Rref
(b) H = 1.5Rref
(c) H = 2.0Rref
(d) H = 2.5Rref
Figure 14. Velocity streamlines on longitudinal plane for quad-rotor hovering at different heights
12 of 19
American Institute of Aeronautics and Astronautics
Velocity streamlines on mid-fuselage longitudinal plane of single-rotor and tandem-rotor are plotted in
figures 11 and 12, respectively. For tilt-rotor, velocity steamlines are plotted on two longitudinal planes,
mid-fuselage plane and mid-rotor plane, in figure 13. For quad-rotor, velocity steamlines are plotted on
three longitudinal plane, mid-fuselage plane, front mid-rotor plane and rear mid-rotor plane, in figure 14.
In the mid-fuselage longitudinal plane, the effect of fuselage is prominent in all the configurations. The
rear rotor wake in the case of a single-rotor is blocked by the fuselage tail, creating a weak and thicker wall jet
in mid-fuselage plane. In tandem-rotor, a strong wake is seen, both in front and rear, as there is no blockage
by the fuselage in these regions, though the fuselage blocks the wake in the fuselage mid section where rotors
overlaps. Since no rotor is placed in mid-fuselage longitudinal plane in both tilt-rotor and quad-rotor, no
strong wakes are observed in this plane. However, due to mixing and interference of two counter-rotating
rotors from port and starboard side, a weak but thicker outwash is observed in fuselage mid plane. At a
lower height, the thickness of this outwash is almost of the order of rotor height. This plays a very important
role in dust particle transport in front of the fuselage for tilt-rotor and quad-rotor.
To compare the dust erosion capability of different rotor configurations, the ground signature (ground
friction velocitu, u∗ ) with oil flow pattern of each configuration at different heights is plotted in figures 15-18.
A significant difference of ground friction velocity pattern has been found between single-rotor and tandemrotor with respect to tilt-rotor and quad-rotor at the same dimensional height. In particular, tandem-rotor
is determined to give strong ground signature for all heights, with the given rotor parameters used.
In all the configurations the ground signature is strong around the perimeter of the rotors. Although
the ground signature gives an idea of dust erosion capability, the transport of dust is entirely governed
by the dynamics of the outwash, as seen by the oil flow pattern. single-rotor configuration creates almost
symmetrical oil flow patterns, suggesting symmetrical dust erosion and transport. In the case of tandemrotor, oil flow patters are converging and parallel on the starboard side of the fuselage, which suggest a
strong dust transport capability on the starboard side. Unlike single-rotor and tandem-rotor, the oil flow
pattern in tilt-rotor and quad-rotor are intense along the fuselage longitudinal axis, which suggest strong
dust transport in the front and rear of fuselage. The shadow of the fuselage in single-rotor, as seen by the
oil flow pattern (tadpole like shape), is observed to be rotating counter-clockwise with height. This is due
to the swirling nature of rotor wake flow. This phenomenon is not prominent in other configurations.
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 15. Ground friction velocity comparison at H = 1.0Rref
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 16. Ground friction velocity comparison at H = 1.5Rref
Dust clouds (iso-surface of 50mg/m3 ) after simulation time of 5s (≈ 15 rotor revolutions) for different
rotor configurations at different heights are shown in Figures 19-22. As seen, the dust cloud engulfs all
rotorcraft configurations at height 1.0Rref after 5s. single-rotor shows almost symmetrical distribution of
dust cloud; whereas, in tandem-rotor dust cloud spread on the sides are dominant. Both tilt-rotor and quadrotor show large cloud formation in the front and back of the fuselage. This same trend is seen at height
1.5Rref , where the dust cloud almost reaches the base of the fuselage in all the configurations. At greater
13 of 19
American Institute of Aeronautics and Astronautics
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 17. Ground friction velocity comparison at H = 2.0Rref
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 18. Ground friction velocity comparison at H = 2.5Rref
height, fuselage in all the configurations are well above the dust clouds. Among all the configurations, at all
rotor heights, the spread and the height of the dust cloud is maximum for the single-rotor and tandem-rotor.
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 19. Dust cloud comparison for multiple rotor configuration at H = 1.0Rref
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 20. Dust cloud comparison for multiple rotor configuration at H = 1.5Rref
To visualize the extent of dust cloud height ahead of the fuselage, as seen from the mid-fuselage lateral
plane, dust cloud on the rear side of mid-fuselage lateral plane is removed. Dust characteristics seen through
mid-fuselage lateral plane (front view) are shown in figures 23-26, where the fuselage longitudinal axis (xaxis) is normal to the images. Although the average dust cloud height around the fuselage sides is greater
for single-rotor and tandem-rotor, at all the heights, the extent of dust cloud height in front of the fuselage
is greater for tilt-rotor and quad-rotor. As seen in the front view at heights 1.0 and 1.5Rref , huge dust
clouds form in front of the tilt-rotor and quad-rotor, thus restricting the pilot’s front visibility. This is
mainly caused by transport of dust particles by stronger and thicker rotor outwash in front of the fuselage
in tilt-rotor and quad-rotor configurations. At lower heights, 1.0 and 1.5Rref , the dust cloud underneath
14 of 19
American Institute of Aeronautics and Astronautics
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 21. Dust cloud comparison for multiple rotor configuration at H = 2.0Rref
(a) Single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 22. Dust cloud comparison for multiple rotor configuration at H = 2.5Rref
the rotors tries to catch up with the fuselage. This phenomenon is intense in single-rotor and tandem-rotor
and is the main cause of dust cloud engulfment of the fuselage. The same phenomenon is observed below
the wings of tilt-rotor and quad-rotor.
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 23. Dust cloud characteristics through fuselage lateral mid plane at H = 1.0Rref
To visualize the extent of dust cloud height on the starboard side of all rotor configurations, as seen from
the mid-fuselage longitudinal plane, dust cloud on the port side of the mid-fuselage longitudinal plane is
removed. Dust cloud as seen through mid-fuselage longitudinal plane (side view) is shown in figure 27-30,
where the y-axis is normal to the plane. It is clearly depicted that the average dust cloud height in front
and back of fuselage is greater for tilt-rotor and quad-rotor. There is huge dust cloud formation, almost to
the height of rotor, on the starboard side of the tandem-rotor at all heights. This is caused by dust carried
by the shifted rotor downwash on the starboard side, as discussed earlier. Catching up of the dust cloud
from underneath the fuselage is also depicted in these plots. In terms of side view perspective, tilt-rotor and
quad-rotor are determined to be better.
15 of 19
American Institute of Aeronautics and Astronautics
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 24. Dust cloud characteristics through fuselage lateral mid plane at H = 1.5Rref
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 25. Dust cloud characteristics through fuselage lateral mid plane at H = 2.0Rref
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 26. Dust cloud characteristics through fuselage lateral mid plane at H = 2.5Rref
IV.
Conclusion
The present study shows that different rotor configurations have different dust cloud formation behavior
during brownout. single-rotor configurations show symmetrical dust cloud evolution around the rotorcraft.
16 of 19
American Institute of Aeronautics and Astronautics
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 27. Dust cloud characteristics through fuselage longitudinal mid plane at H = 1.0Rref
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 28. Dust cloud characteristics through fuselage longitudinal mid plane at H = 1.5Rref
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 29. Dust cloud characteristics through fuselage longitudinal mid plane at H = 2.0Rref
In tandem-rotor, the dust cloud evolution is significant along the sides of the fuselages. Tilt-rotor and
quad-rotor shows substantial dust cloud formation on the front and rear of fuselage. Although the amount
of dust entrainment in tandem-rotor is high, from the pilot’s front view perspective, the visibility is better
17 of 19
American Institute of Aeronautics and Astronautics
(a) single-rotor
(b) Tandem-rotor
(c) Tilt-rotor
(d) Quad-rotor
Figure 30. Dust cloud characteristics through fuselage longitudinal mid plane at H = 2.5Rref
compared to other configurations. Both tilt-rotor and quad-rotor show lover visibility because of large dust
cloud evolution in front of the fuselage.
Acknowledgments
Authors would like to thanks Air Force Office of Scientific Research (AFOSR) for the grant on the Multi
University Research Initiative on “Rotorcraft Brownout: Advanced Understanding, Control and Mitigation”.
Authors express their gratitude to Dr. James D. Iversen for his suggestion and guidance during the course
of the project. In addition, the authors appreciate all the help and discussion with Dr. William Warmbrodt
and Dr. Marvin Moulton during the course of study. Authors also like to thanks Department of Aerospace
Engineering, Iowa State University for providing the computational resources.
References
1 Rogers. S. J., “Evaluation of The Dust Cloud Generated by Helicopter Rotor Downwash,” Technical Report 67-81, U.S.
Army Aviation Material Laboratories, 1968.
2 Chatten, C., “Sandblaster 2 Support of See-Through Technologies for Particulate Brownout”, Task 5 Technical Report,U.S. Army Aviation and Missile Command, Midwest Research Institute (MRI), 2007.
3 Keller, J. D., Whitehouse, G. R., Wachspress, D. A. ,Teske, M. E., and Quackenbush, T. R. “A Physics-Based Model of
Rotorcraft Brownout for Flight Simulation”, 62nd Annual American Helicopter Society Forum, Phoenix, AZ, 2006.
4 Ryerson, C. C., Hachnel, R. B., Koenig, G. G., and Moulton, M. A., “Visibility Enhancement in Rotorwash Clouds”,
AIAA Aerospace Sciences Meeting and Exhibit Reno, Nevada, 2005.
5 Haehnel, R. B., Moulton, M. A., Wenren, Y., and Steinhoff, J., “A Model to Simulate Rotorcraft-Induced Brownout”,
64th Annual Forum of the American Helicopter Society Montreal, 2008.
6 Phillips, C. and Brown, R. E., “Eulerian simulation of the fluid dynamics of helicopter brownout”, American Helicopter
Society 64th Annual Forum, Montreal, Canada, April 29 - May 1, 2008.
7 Phillips, C., Kim, H. W. and Brown, R. E., “The Effect of Rotor Design on the Fluid Dynamics of Helicopter Brownout”,
35th European Rotorcraft Forum Germany, 2009.
8 Wachspress, D. A., Whitehouse, G. R., Keller, J. D., Yu, K., Gilmore, P., Dorsett, M., and McClure, K., “A high fidelity
brownout model for real time flight simulation and trainer”, American Helicopter Society 65th Annual Forum, Grapevine, 2009.
9 Keller, J. D., Whitehouse, G. R., Wachspress, D. A., Milton, E. T., Quackenbush, T. R., “A Physics-Based Model of
Rotorcraft Brownout for Flight Simulation Applications”, American Helicopter Society 62nd Annual Forum, Phoenix, AZ,
2006.
10 Haehnel, R. and Dade, W. B., “Physics of Particle Entrainment Under the Influence of an Impinging Jet”, Proceedings
Army Science Conference, Orlando, Florida, 2008.
11 Rajagopalan, R. G. and Fanucci, J. B., “Finite difference model for vertical axis wind turbines”, Journal of Propulsion
and Power, Vol. 1, 1985, pp. 432–436.
12 Shih, T. H., Liou, W. W., Shabbir, A., Yang, Z. and Zhu, J., “A new eddy-viscosity model for high Reynolds number
turbulent flows–Model development and validation”, Computers Fluids, Vol. 24, No. 3, 1995, pp. 227–238.
13 Kato, M. and Launder, B. E., “he Modeling of Turbulent Flow Around Stationary and Vibrating Square Cylinders”,
Proc. 9th Symposium on Turbulent Shear Flows, Kyoto 10.4.1–10.4.6, 1993.
18 of 19
American Institute of Aeronautics and Astronautics
14 Rajagopalan, R. G. and Mathur, S, R., “Three Dimensional Analysis of a Rotor in Forward Flight”, Journal of American
Helicopter Society, Vol. 38, 1993, pp. 14–25.
15 Crowe, C., Sommerfeld, M. and Tsuji, Y., Multiphase flows with droplets and particles. CRC Press (Boca Raton, Fla),
1998.
16 Shirolkara, J. S., Coimbrab, C. F. M. and Queiroz McQuay, M., “Fundamental aspects of modeling turbulent particle
dispersion in dilute flows”, Progress in Energy and Combustion Science, Vol. 22, No. 4, 1984, pp. 363–399.
17 Lu, H. and Shao, Y., “Towards quantitative prediction of dust storms: An integrated wind erosion modelling system and
its application”, Environmental Modelling and Software, Vol. 16, No. 3, 2001, pp. 233–249.
18 Nalpanis, P., Hunt, J. C. R.and Barrett, C. F., “Saltating particles over flat beds”, Journal of Fluid Mechanics, Vol. 251,
1993, pp. 661–685.
19 Marticorena, B. and Bergametti, G., “Modeling the atmospheric dust cycle: 1. Design of a soil-derived dust emission
scheme”, Journal of Geophysical Research, Vol. 100, No. D8, 1995, pp. 415–430.
20 Bagnold, R. A., The physics of blown sand and desert dunes, Journal of Atmospheric Science, London: Methuen, 265
pp, 1941.
21 Manninen, M., Taivassalo, V., and Kallio, S., “On the mixture model for multiphase flow”, VTT Publications 288,
Technical Research Centre of Finland, 1996.
22 Fluent 6.3, Fluent User’s Guide, ANSYS, Inc., 2006.
23 Patankar, S. V., Numerical heat transfer and fluid flow, Washington, DC, Hemisphere Publishing Corp., 1980.
24 Rabbott, J. P., “Static-Thrust Measurements of the Aerodynamic Loading on a Helicopter Rotor Blade”, Technical
Report, National Aeronautics and Space Administration, Washington DC, 1956.
25 Nathan, N. D. and Green, R. B., “Measurements of a rotor flow in ground effect and visualization of the brown-out
phenomenon”, Helicopter Society Annual Forum, Montreal, Canada, 2008.
26 Wilcox, D.C., Turbulence Modeling for CFD, 2nd Ed., DCW Industries, Inc., 1998.
27 Tennekes, H. and Lumley, J. L., A First Course in Turbulence, MIT Press, Cambridge, MA, 1972.
28 Launder, B. E., Reece, G. J. and Rodi, W., “Progress in the Development of a Reynolds-Stress Turbulent Closure”,
Journal of Fluid Mechanics, Vol. 68, No. 3, 1975, pp. 537–566.
29 Launder, B. E. and Spalding, D. B., “The numerical computation of turbulent flows”, Computer Methods in Applied
Mechanics and Engineering, Vol. 3, No. 2, 1974, pp. 269–289.
30 Jones, W. P. and Launder, B. E., “The Prediction of Laminarization with a Two-Equation Model of Turbulence”,
International Journal of Heat and Mass Transfer, Vol. 15, 1972, pp. 301–314.
31 Shao, Y., Physics and modelling of wind erosion, Boston : Kluwer Academic, Dordrecht, 2000.
32 Owen, P. R., “Saltation of uniform grains in air”, Journal of Fluid Mechanics, Vol. 20, No. 2, 1964, pp. 225–242.
33 Iversen, J. D., Greeley, R. and Pollack, J. B., “Wind blown dust on Earth, Mars and Venus”, Journal of Atmospheric
Science, Vol. 33, 1976, pp. 2425–2429.
34 Greeley, R. and Iversen, J. D., “Wind as a Geological Process”, Cambridge University Press, 1985, pp. 99–100.
35 Gillette, D. A., “Fine particulate emissions due to wind erosion”, Transactions of the American Society of Agricultural
Engineers, Vol. 20, 1977, pp. 890–897.
36 Durst, F., Miloevic, D. and Schnung, B., “Eulerian and Lagrangian predictions of particulate two-phase flows: a numerical
study”, Journal of Instructional Computing Research, Vol. 8, No. 2, 1984, pp. 101–115.
37 Crowe, C. T., Troutt, T. R. and Chung, J. N., “Numerical models for two-phase turbulent flows”, Annual Review of
Fluid Mechanics, Vol. 28, 1984, pp. 11–43.
38 Ishii, M., Thermo-fluid dynamic theory of two-phase flow, Paris: Evrolles, 1975.
39 Craft, T. J., Graham, L. J. W., and Launder, B. E., “Impinging jet studies for turbulence model assessment. Part 2: An
examination of the performance of four turbulence models”, International Journal of Heat and Mass Transfer, Vol. 36, 1993,
pp. 2685–2697.
40 Preston, J. R., “VTOL Downwash/Outwash Operational Effects Model”, Annual Forum Proceedings, American Helicopter Society, Vol. 50, No. V2, 1994, pp. 1101.
41 Phillips, C. and Brown, R. E., “The effect of helicopter configuration on the fluid dynamics of brownout”, 34th European
Rotorcraft Forum, 16th - 19th Sept 2008, Liverpool, England.
19 of 19
American Institute of Aeronautics and Astronautics