FRICTION - From The Virginia Tech Aerodynamics and Design Software Collection

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

Program FRICTION
FRICTION provides an estimate of laminar and turbulent skin friction and form drag suitable for
use in aircraft preliminary design. The program has its roots in a program by Ron Hendrickson at
Grumman. It runs on any computer. The input requires geometric information and either the
Mach and altitude combination, or the Mach and Reynolds number at which the results are
desired. It uses standard flat plate skin friction formulas and form factors.

The method uses a build-up approach, where the parasite drag for each component is found and
then summed to find the total parasite drag. The output provides the contribution for each
component. The basic formula is

N FFj CFj Swet j


C D0 = ∑
j =1 Sref

where N is the number of components used to model the configuration. The form factor, FF,
depends on whether the component is modeled as a planar surface or a body of revolution. The
formulas for the form factor are given below. The skin friction coefficient, CF, can be for
either laminar or turbulent flow and are also described below.

At component junctures, an interference drag can be added, and beyond drag rise the wave drag
must be added. See the Gur, Mason, and Schetz, “Full Configuration Drag Estimation,” Journal
of Aircraft, July-Aug. 2010, pp. 1356-1367, for more details.

The compressibility effects on skin friction are found using the Eckert Reference Temperature
method for laminar flow and the van Driest II formula for turbulent flow. The basic formulas are
valid from subsonic to hypersonic speeds, but the implementation makes assumptions that limit
the validity to moderate supersonic speeds (about Mach 3). The key assumption is that the
vehicle surface is at the adiabatic wall temperature (the user can easily modify this assumption).
As noted above, form factors are used to estimate the effect of thickness on drag, and a
composite formula is used to include the effect of a partial run of laminar flow.
Laminar flow
The Blasius formula for skin friction is used, adjusted for compressibility using the Eckert
Reference Temperature Method. This particular version is the one given by F.M. White in
Viscous Fluid Flow, McGraw-Hill, New York, 1974, pp. 589-590. In this method the

2/6/15 1
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

incompressible skin friction formula is used, with the fluid properties chosen at a specified
reference temperature, which includes both Mach number and wall temperature effects.

First, assumptions are made for the fluid properties:∗ Prandtl number, Pr = 0.72, Recovery
factor, r = Pr1/2, specific heat ratio, γ = 1.4, and edge temperature, Te = 390 (°R). Then, for a
given edge Mach number, Me, and ratio of wall temperature to adiabatic wall temperature
TW/TAW; compute:
TW T ⎛ γ −1 2⎞
= W ⎜1 + r Me ⎟ .
Te TAW ⎝ 2 ⎠
Remember that
⎛ γ −1 2⎞
TAW = Te ⎜ 1 + r Me ⎟
⎝ 2 ⎠
and then compute the reference temperature:
T* ⎛T ⎞
≅ .5 + .039M e2 + 0.5 ⎜ w ⎟
Te ⎝ Te ⎠
The Chapman-Rubesin constant based on the reference temperature and Sutherland’s viscosity
law is then computed from:
1/2
⎛T*⎞ ⎛ 1 + K / Te ⎞
C =⎜ ⎟
*
⎜⎝ T * / T + K / T ⎟⎠
⎝T ⎠ e e e

where K = 200°R for air.


Finally, the local friction coefficient (τw/q) is found from the standard Blasius formula, with
C* added,

.664 C *
Cf =
Re x
and
CF = 2C f
which comes from
x′= x
F 1
qx x x ′∫= 0
CF = = C f ( x ′ )dx ′

Recall that CF accounts for one side of the plate only, so that if both sides are required for a drag
estimate, then the skin friction coefficient, CD, is twice CF because the reference area is based on
one side only, i.e., Sref ≈ 1/2 Swet.


These values can be changed easily in the source code.

2/6/15 2
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

Note that the results are not sensitive to the value of edge temperature for low Mach
numbers, and therefore, an exact specification of Te is not required. This method is implemented
in subroutine lamcf.

Turbulent flow
For turbulent flow the so-called van Driest II Method is employed. This method was selected
based on the recommendation of E.J. Hopkins and M. Inouye, contained in “An Evaluation of
Theories for Predicting Turbulent Skin Friction and Heat Transfer on Flat Plates at Supersonic
and Hypersonic Mach Numbers,” AIAA J., Vol. 9, No. 6, June 1971, pp. 993-1003. The
particular algorithm is taken from NASA TN D-6945, “Charts for Predicting Turbulent Skin
Friction From the Van Driest Method (II),” also by E.J. Hopkins, and dated October 1972.
Again, assumptions are made for the fluid properties: turbulent flow recovery factor, r = .88,
specific heat ratio, γ = 1.4, and edge temperature, Te = 222 (°K). Then, for a given edge Mach

number, Me, and ratio of wall temperature to adiabatic wall temperature TW/TAW the calculation
is started by computing the following constants:
γ −1 2
m= Me
2

Tw T T
F= = w ⋅ AW
Te TAW Te
where
TAW
= 1 + rm
Te
Tw = F ⋅ Te
1/2
⎛ rm ⎞
A=⎜ ⎟
⎝ F⎠
1 + rm − F
B=
F
2A − B
2
α=
( )
1/2
4A 2 + B 2
B
β=
( 4A )
1/2
2
+ B2

2/6/15 3
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

rm
Fc = M e > 0.1
( )
2
sin −1 α + sin −1 β
2
⎛1+ F ⎞
=⎜ ⎟ M e ≤ 0.1
⎝ 2 ⎠
and
⎛ 122 ⎞
1+ × 10 −5 /Tw
µe 1 ⎜ Tw ⎟
Fθ = = ⎜ ⎟
µw F ⎜ 1 + 122 × 10 −5 /Te ⎟
⎝ Te ⎠
which is the Keyes viscosity law.
Finally,

Fx =
Fc
The analysis proceeds using barred quantities to denote “incompressible” variables, which are
intermediate variables not used except to obtain the final results. Given the Reynolds number,
Rex, an iteration is used to obtain the final results. Proceed as follows, finding

Re x = Fx Re x
solve the Kármán-Schoenherr turbulent skin friction formula

= log ( Re x CF )
.242
CF
for CF . Use as an initial guess
.074
CF0 = .
Re.20
x

Then, Newton’s method is applied to the problem:

f ( CF ) = 0 ⇒ CFi+1 = CFi −
f
f′
which becomes for this equation:

i +1

= C ⎢1 +
{
.242 − CFi log Re x CFi ( )} ⎤⎥
{ }
i
C F F
⎢ .121 + CFi / ln10 ⎥
⎢⎣ ⎥⎦

Once this iteration is completed, and CF is known,

CF
CF =
Fc

2/6/15 4
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

Note that this value applies to one side of a plate only, so it must be doubled if the friction on
both sides is desired to account for the proper reference areas. Here again, the results are not
sensitive to the value of edge temperature for low Mach numbers, and the default value should
be adequate for most cases. This formula is implemented in routine turbcf.

Composite formula
When the flow is laminar and then transitions to turbulent, an estimate of the skin friction is
available from a composite of the laminar and turbulent skin friction formulas using
Schlichting’s formula (see T. Cebeci and P. Bradshaw, Momentum Transfer in Boundary Layers,
McGraw-Hill, New York, 1977, pp. 187). Given the transition position, xc/L and ReL, compute

⎛x ⎞
Re c = ⎜ c ⎟ Re L
⎝ L⎠
and compute the laminar skin friction based on Rec and the turbulent skin friction twice, based
on both Reynolds numbers and then find the value that includes both laminar and turbulent flow
from:
⎛x ⎞
CF = CFTURB (Re L ) − ⎜ c ⎟ ⎡⎣CFTURB (Re c ) − CFLAM ( Re c ) ⎤⎦
⎝ L⎠
Several formulas are available, are all roughly equivalent, and have been evaluated extensively
for incompressible flow. They are only approximate for compressible flow.
Form factors
To include the effects of thickness, it has been found that the skin friction formulas should be
adjusted through the use of form factors. Two different factors are used in this code. For wing-
like shapes,
4
⎛t⎞ ⎛t⎞
FF = 1.0 + 2.7 ⎜ ⎟ + 100 ⎜ ⎟
⎝ c⎠ ⎝ c⎠

where t/c is the thickness ratio of particular component.∗ For bodies,


1.5 3
⎛ d⎞ ⎛ d⎞
FF = 1.0 + 1.5 ⎜ ⎟ + 50 ⎜ ⎟
⎝l⎠ ⎝l⎠
where d/l is the ratio of diameter to length. This is the reciprocal of the fineness ratio. To model a
nacelle remove the inlet area.


The factor is from Torenbeek, Synthesis of Subsonic Airplane Design, Delft University Press, 1982, based on a
comparison of FRICTION results with the parasite drag of an NACA 0012 airfoil.

2/6/15 5
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

Program Operation:
Running the program, you will be prompted for the name of an input data set. The maximum
name length is 15 characters. The output is sent to the screen, but can be sent to a file by
changing the value of IWRIT to something other than 6 in the main program. The sample data
case on the disk is F15frict.inp. Note that the units are assumed to be in the English system.

Notes about units in FRICTION (by Prof Todd Lowe when he was our assistant)
FRICTION exists in two forms (FORTRAN executable and MATLAB script) with two methods
for the trial runs (Altitude and Reynolds number per length). Proper units must be used to obtain
the correct drag coefficient. The units that may be used differ depending on the trial run method
and whether the FORTRAN or MATLAB version is used.

FORTRAN executable:
- If inputting the altitude, all lengths must be in feet, and all areas must be in ft2. The
FORTRAN program assumes that the altitude is entered in thousands of feet.
- If using Reynolds number per length, any consistent set of units may be used. However,
the Reynolds number per unit length must be entered as the Reynolds number divided by
the characteristic length in the same units. For example, if the configuration dimensions
are entered in mm and mm2, the Reynolds number per unit length must be the Reynolds
number divided by the configuration characteristic length entered in mm.

MATLAB:
- For altitude input, all configuration data must be entered in either ft/ft2 or m/m2. After
choosing the altitude method for trial runs, a prompt will appear asking the user to choose
metric (enter 0) or English (enter 1) units. These units must be consistent with the
configuration data input If metric units are chosen, the altitude must be entered in km.
For English units, the altitude must be entered in thousands of feet.

- For Reynolds number per unit length, the MATLAB script behaves exactly like the
FORTRAN executable in this case. Any consistent set of units may be used. The
Reynolds number per unit length is the Reynolds number divided by the characteristic
length in the units used in the configuration data.

2/6/15 6
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

INPUT
Card Field Columns Variable Description
1 1 1-60 Title Card
2 1 1-10 SREF Full Scale reference Area
2 11-20 SCALE 1./SCALE, i.e. 1/10 scale is input as 10.
3 21-30 FNCOMP number of component cards to be read in (15 max).
4 31-41 FINMD input mode: = 0.0, input Mach and altitude
= 1.0, input Mach and Reynolds No.
per unit length
Card Field Columns Variable Description
3 1 1-16 COMP(I) Component Name
2 21-30 SWET(I) Wetted Area (i.e., top and bottom sides of the wing,
and both left and right sides, the total area that is
exposed to the air)
3 31-40 REFL Reference Length, i.e. the average chord for a wing.
4 41-50 TC(I) t/c for planar surf. or d/l (1/F) for body of revolution
5 51-60 FICODE Component type clue
= 0.: Planar surface
= 1.: Body of revolution
6 61-70 FTRANS Transition location
= 0. : means boundary layer is all turbulent
= 1. : " " " " " laminar.
values between 0 and 1 approximate the value of
the friction of the laminar/turbulent boundary layer
at the specified length fraction of the component.
Note: card 3 is repeated NCOMP times
Card Field Columns Variable Description
4 1 1-10 XME Mach number
2 11-20 XINPUT if FINMD = 0.0, this is the Altitude (in 1000 feet)
if FINMD = 1.0, this is the Reynolds no. per unit length
in millions
Note: Card 4 is repeated for each value of Mach and altitude desired. The program stops when
either the end of the data is reached or a Mach number of zero is read.
Output: The input is echoed to allow for easy check of data and to keep all information together.
Then the drag calculation for each M,h or M,Re/L is made. First, the reference areas, lengths,
thicknesses, form factors and the transition position are output. These values are fixed for each
combination of Mach and Reynolds number. Next, for each case the Reynolds number of each
component and the basic skin friction are found. Then the skin friction times the wetted area and
the skin friction times the wetted area and form factor are found. Finally, the latter is divided by
the reference area and the contribution to the total drag in terms of a drag coefficient for the
particular component, CDCOMP, is then found. These columns are summed, and the bottom
value under the CDCOMP column is the total skin friction and form drag coefficient. After all
the conditions are computed, a summary of results is presented as a table at the end of the output.

2/6/15 7
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

Sample input for program FRICTION:


F - 15 AIRCRAFT
608. 1. 7. 0.0
FUSELAGE 550.00 54.65 .05500 1.0 0.0
CANOPY 75.00 15.0 .12000 1.0 0.0
NACELLE 600.00 35.0 .04000 1.0 0.0
GLV/SPONSON 305.00 35.5 .117 1.0 0.0
OUTB'D WING 698.00 12.7 .05000 0.0 0.0
HORIZ. TAIL 222.00 8.3 .05000 0.0 0.0
TWIN V. T. 250.00 6.7 .0450 0.0 0.0
0.200 35.000
1.200 35.000
2.000 35.000
0.000 0.000
Sample output from program FRICTION:

Enter name of data set:


F15frict.inp

FRICTION - Skin Friction and Form Drag Program


W.H. Mason, Department of Aerospace and Ocean Engineering
Virginia Tech, Blacksburg, VA 24061 email:[email protected]
version: Jan. 28, 2006

CASE TITLE: F - 15 AIRCRAFT

SREF = 608.00000 MODEL SCALE = 1.000 NO. OF COMPONENTS = 7


input mode = 0 (mode=0: input M,h; mode=1: input M, Re/L)

COMPONENT TITLE SWET (FT2) REFL(FT) TC ICODE FRM FCTR FTRANS


FUSELAGE 550.0000 54.650 0.055 1 1.0205 0.0000
CANOPY 75.0000 15.000 0.120 1 1.0744 0.0000
NACELLE 600.0000 35.000 0.040 1 1.0124 0.0000
GLV/SPONSON 305.0000 35.500 0.117 1 1.0712 0.0000
OUTB'D WING 698.0000 12.700 0.050 0 1.1356 0.0000
HORIZ. TAIL 222.0000 8.300 0.050 0 1.1356 0.0000
TWIN V. T. 250.0000 6.700 0.045 0 1.1219 0.0000

TOTAL SWET = 2700.0000

REYNOLDS NO./FT =0.480E+06 Altitude = 35000.00 XME = 0.200

COMPONENT RN CF CF*SWET CF*SWET*FF CDCOMP


FUSELAGE 0.262E+08 0.00251 1.38212 1.41047 0.00232
CANOPY 0.720E+07 0.00309 0.23164 0.24889 0.00041
NACELLE 0.168E+08 0.00269 1.61561 1.63573 0.00269
GLV/SPONSON 0.170E+08 0.00269 0.81944 0.87782 0.00144
OUTB'D WING 0.609E+07 0.00318 2.21681 2.51746 0.00414
HORIZ. TAIL 0.398E+07 0.00342 0.75829 0.86114 0.00142
TWIN V. T. 0.321E+07 0.00355 0.88656 0.99464 0.00164
SUM = 7.91048 8.54615 0.01406

FRICTION DRAG: CDF = 0.01301 FORM DRAG: CDFORM = 0.00105

REYNOLDS NO./FT =0.288E+07 Altitude = 35000.00 XME = 1.200

2/6/15 8
FRICTION - from the Virginia Tech Aerodynamics and Design Software Collection

COMPONENT RN CF CF*SWET CF*SWET*FF CDCOMP


FUSELAGE 0.157E+09 0.00175 0.96201 0.98175 0.00161
CANOPY 0.432E+08 0.00211 0.15826 0.17004 0.00028
NACELLE 0.101E+09 0.00186 1.11769 1.13160 0.00186
GLV/SPONSON 0.102E+09 0.00186 0.56700 0.60740 0.00100
OUTB'D WING 0.366E+08 0.00216 1.51055 1.71542 0.00282
HORIZ. TAIL 0.239E+08 0.00231 0.51314 0.58274 0.00096
TWIN V. T. 0.193E+08 0.00239 0.59777 0.67064 0.00110
SUM = 5.42643 5.85959 0.00964

FRICTION DRAG: CDF = 0.00893 FORM DRAG: CDFORM = 0.00071

REYNOLDS NO./FT =0.480E+07 Altitude = 35000.00 XME = 2.000

COMPONENT RN CF CF*SWET CF*SWET*FF CDCOMP


FUSELAGE 0.262E+09 0.00140 0.76912 0.78490 0.00129
CANOPY 0.720E+08 0.00169 0.12643 0.13585 0.00022
NACELLE 0.168E+09 0.00149 0.89337 0.90449 0.00149
GLV/SPONSON 0.170E+09 0.00149 0.45321 0.48550 0.00080
OUTB'D WING 0.609E+08 0.00173 1.20667 1.37032 0.00225
HORIZ. TAIL 0.398E+08 0.00185 0.40980 0.46538 0.00077
TWIN V. T. 0.321E+08 0.00191 0.47731 0.53550 0.00088
SUM = 4.33591 4.68193 0.00770

FRICTION DRAG: CDF = 0.00713 FORM DRAG: CDFORM = 0.00057

SUMMARY

J XME Altitude RE/FT CDF CDFORM CDF+CDFORM


1 0.200 0.350E+05 0.480E+06 0.01301 0.00105 0.01406
2 1.200 0.350E+05 0.288E+07 0.00893 0.00071 0.00964
3 2.000 0.350E+05 0.480E+07 0.00713 0.00057 0.00770

END OF CASE

Press RETURN to quit the program.

2/6/15 9

You might also like