Classical Inflow Models
Background
• In Part II and Part III of our lecture
notes, we used Glauert’s inflow model to
compute the induced velocity in forward
flight.
• This model gives acceptable results for
performance, but does not give good
results for blade dynamics or vibratory
loads.
• It fails miserably for Blade-Vortex-
Interactions.
T 2 Av V cos V sin v
2 2
Why does the model fail in forward flight?
Because most of the wake is the first and
second quadrants.
We can not expect the induced velocities to
be uniform!
Freestream
Tip Vortices
Rotor Disk
Experiment by Harris
• Because most of the tip vortices are in the first and
fourth quadrants, the induced velocity associated with
these vorticies (i.e. the inflow) is very high in the first and
fourth quadrants.
• The resultant reduced angle of attack leads to smaller lift
forces in the aft region, compared to uniform inflow
model.
• The blade responds to this reduced loads 90 degrees
later.
• It flaps down (leading to a more negative b1s) more than
expected.
• Harris experimentally observed this, in a paper published
in 1972 in the Journal of the American Helicopter Society
(See Wayne Johnson, page 274, figure 5-39).
Harris Wind Tunnel Test
Harris Wind Tunnel Test
H-34 Flight Test (Home Work 3)
SA-349 Flight Test
Effects of Non-Uniform Inflow on
Blade-Vortex-Interaction Loads
• During descent the tip vortices are pushed up
against the rotor disk by the freestream, leading
to a very close spacing between the blades and
the vortices.
• According to Biot-Savart law, this leads to very
high, and very rapidly varying, induced velocity.
• This affects the airloads dramatically.
• Ignoring this rapid variation in the inflow and
using a Glauert inflow model will lead to a
severe underprediction of vibratory loads, and
associated aerodynamically generated noise.
Blade Vortex Interaction
The velocity field experienced by the blade changes dramatically in a matter of
mili-seconds!
Efforts to Improve Inflow Models
• Engineers and researchers recognized
very quickly that there is need for
improvements in the inflow model.
• Until the mid 1960s, the work was
analytical or semi-empirical. This approach
is called classical vortex theory. See
pages 134-141 of text, and our web site.
• Starting 1960s, numerical approaches
based on Biot-Savart law became popular.
Coleman’s Model
Coleman divided the
Real Tip Vortex is helical, Continuous helical vortex into a
and distorted. Its strength varies Series of circular rings.
along its length. He assumed that the
vortex strength is constant.
Coleman used Biot-Savart
law, and numerical
integration to compute the
Side View
Induced velocity at a few Induced velocity v + Normal component of freestream
points on the rotor disk. = WR
Rotor Disk
Edgewise component of
Freestream Velocity = WR Successive Rings of Vortices
v v Glauert 1 k x r cos
where ,
1 1
k x tan G=constant in Coleman’s Model
2 2
Castles’ et al’s Model
• This model is discussed in NACA report
1184. A pdf file may be found at the web
site.
• Castles replaced the numerical integration
in Coleman’s model with analytical
integration.
• These authors also replaced the individual
rings by a continuous sheet of vorticity.
Castles’ Model
The individual rings were replaced by a skewed cylindrical surface.
Vorticity strength was uniform on the cylinder surface.
Biot-Savart Law was used. Results are supplied as charts and graphs.
Heyson et al’s Model
• This model is discussed in the NACA
Report 1319. An electronic version may be
found at our class site.
• This model rectifies one of the
assumptions in Castles’ and Coleman’s
models, namely that the rotor only sheds
tip vortices.
Radial Variation of Loads causes a
continuous shedding of vorticity
Heyson assumed that the
Circulation G varies linearly
With radius.
r/R
Heyson modeled each of the trailing vortices as a vortex sheet whose shape
is in the form of a skewed cylinder. The effects of each of the cylinders may
be superposed, using the analytical formulas developed by Castles’ et al.
Heyson’s Model
At each radial location
A vortex sheet of shape
Similar to a circular
Cylinder is shed.
The strength is constant,
Both along the axis of the
Cylinder and around the azimuth.
Advanced Inflow Models
Background
• Around 1965, more powerful digital
computers became readily available.
• Engineers began to model the tip vortex
as a skewed, distorted, helix.
• The strength of the tip vortex was allowed
to vary with azimuth, acknowledging the
fact that the blade loading changes with
the azimuthal position of the blade.
Background (Continued)
• Sadler at Bell Helicopter, and Scully at MIT developed
some of the earliest techniques.
• Sadler used a rigid helical wake, while Scully allowed for
the wake to deform due to self-induced velocities.
• The induced velocity at the “strips” on the blade were
computed using Biot-Savart Law.
• Modern methods (e.g. CAMRAD-II) not only model the
tip vortex, but also the inboard vortices, and shed
vortices.
• See our web site for several publications related to
CAMRAD-II.
Shed Wake
When the lift (or bound circulation) around an airfoil changes with time,
Circulation that is equal in magnitude to the bound circulation but opposite
In strength is shed into the wake.
Trailing and Shed Wake
Representation
The blade is divided into a number of strips, as in HW#1
Hub and 2
Strip j Tip
The end points of
Shed Vortex vortex segments are
Called markers.
Tip Vortex
Inboard
Trailing Vortex
Trialing Vortex Strength
Bound Circulation Bound Circulation
On this strip G On next strip GdG
dG
Right hand screw rule is applied to assign the orientation of the vortex.
The strength of the most recently shed trailing vortex segment may be
thought of as the change in the bound vortex strength between adjacent “strips.”
Shed Vortex Strength
Strip j Blade position at
Gjn+1 next time step
n+1
Tip vortex
Blade position
At time step n
Strength of shed
Vortex segment = Gjn+1-Gjn
The strength of the shed vortex segments
Will vary from one strip to the next.
Induced Velocity Calculation
• The calculations are done in a time marching
mode.
• The strength of the shed and trailing vortices are
known from the previous time steps.
• The geometry of the wake is assumed to be a
helix (rigid), prescribed (deformed wake, curve
fitted from experiments) or free wake.
• Free wake geometry is computed by allowing
the wake to move at the freestream velocity plus
the induced velocity computed at the junction
points (markers).
Rigid Wake Model
x r
cos( f ) xf xf
R R
y r
sin( f ) yf yf
R R
z z0
zf zf
R R
where the is the azimuthal angle of the reference blade,
and f is the vortex wake age. x, x and z are advance ratio
Components along x-, y-, and z- directions.
The Glauert uniform inflow model is used to estimate the
inflow components x, y and z , the three components of .
cT
tan i tan
2 2 2
Wake Age
f
Tip Vortex
x r
cos( f ) xf xf
R R
y r
sin( f ) yf yf
R R
z z0
zf zf
R R
x
Prescribed Wake Model
In this model, the x and y coordinates of the tip vortex are prescribed
from a rigid wake model. The vertical displacements of the tip vortices
are given as:
z
f EG
V
i
R
where E is an envelop function given by:
E A0f exp( A1f ) if f 4
E Mf B if f 4
N
G C n cos nf Dn sin nf
n 0
Here, A0, A1, M, B, Cn and Dn are all empirical constants listed in
“Egolf, A. and Landgrebe, A. J., “Helicopter Rotor Wake Geometry
and Its Influence in Forward Flight, Vol. I, NASA CR-3726.”
Free Wake Geometry
Gt-Hybrid freewake model
