NASA Technical Reports Server (NTRS) 20010043665: Prediction of Business Jet Airloads Using The Overflow Navier-Stokes Code

10 Pages·0.7 MB·English

by  NASA Technical Reports Server (NTRS)

• 1
AMU_-2001-1004
PREDICTION OF BUSINESS JET AIRLOADS USING THE
OVERFLOW NAVIER-STOKES CODE
Elias Bounajem"
Cessna AircraftCompany, Wichita, Kansas 67215
and
Pieter G. Buning"
NASA Langley Research Center, Hampton, Virginia 23681
Abstract regular basis to support design efforts in
such areas as aircraft design, propulsion
The objective of this work is to evaluate the system design and integration, combustion,
application of Navier-Stokes computational ship design and the automotive industry, to
fluid dynamics technology, for the purpose of name afew.
predicting off-design condition airloads on a
business jet configuration inthe transonic The aircraft industry is a primary customer of
regime. The NASA Navier-Stokes flow CFD technology. Depending onthe task at
solver OVERFLOW with Chimera overset hand, the aircraft designer has awide range
grid capability, availability of several of sophistication to chose from. This range
numerical schemes and convergence includes: full modeling of viscous effects as
acceleration techniques was selected for this available in Navier-Stokes type codes; Euler
work. A set of scripts which have been codes, transonic small disturbances and full
compiled to reduce the time required for the potential codes when viscous effects are not
grid generation process are described. of primary influence; and finally, linearized
Several turbulence models are evaluated in potential flow codes for shock free flows or
the presence of separated flow regions on when only an approximate answer is being
the wing. Computed results are compared to sought. These codes are implemented with
available wind tunnel data for two Mach different types of grids: multiblock, patched
numbers and a range of angles-of-attack. or overset grids if astructured grid approach
Comparisons of wing surface pressure from isconsidered, Cartesian and unstructured
numerical simulation and wind tunnel grids for unstructured methods. The
measurements show good agreement upto literature is rich with publications identifying
fairly high angles-of-attack. the pros and cons of each of these methods.
Introduction The bulk of CFD work inthe aircraft design
arena has been inthe cruise flight regime.
Under these conditions, the airflow is still
Computational Fluid Dynamics (CFD)
attached to the surface withsmall regions of
technology has seen remarkable advances
separated flow, and CFD calculations are
known to be quite reliable inpredicting
*Member, AIAA
"*Associate Fellow, AIAA. aircraft performance characteristics.
Copyright ©2001 bythe American InstituteofAeronautics and
Astronautics, Inc. All rights reserved.
inthe last few years. CFD is used today on a Areas which have not received as much
attention are those that deviate from the
1
American Institute ofAeronautics and Astronautics
AIAA-2001-1004
cruise condition. These situations are components. Five components were
encountered when the entire flight envelope identified: wing, fuselage, pylon, nacelle and
of the airplane is of interest. Here, wing-body fairing. For the wing, pylon, and
combinations of angle-of-attack, Mach and nacelle, the CATIA CAD model was
Reynolds number can bechallenging for any imported in IGES format into the NASA
CFD code. High angle-of-attack will cause Langley GridTool software [2], where a
the flow to separate. As the high subsonic surface grid was generated. The fuselage
regime is approached, the flow will have surface grid was generated directly in CATIA
numerous shock waves on the wing, pylon, to improve surface grid quality resulting from
etc. The interaction of shock waves with the projection onto the original geometry. As for
boundary layer complicates flow patterns the wing-body fairing (Figure 2), it was
and makes CFD prediction significantly more deemed appropriate to represent it with a
difficult. With the presence of large combination of collar grids and therefore no
separated flow regions, the focus shifts more computational surface grid was generated
and more towards turbulence modeling. for this component.
Here, the validity of the model's underlying
assumptions becomes critical to the quality Collar Grids
of the solution the CFD code can provide.
The next step inthe process is to make use
The present work aims at evaluating the of the collar grid approach [3] inthe overlap
applicability of Navier-Stokes CFD region between the various components.
technology to predict off-design airloads on a This approach is used to ensure good quality
business jet configuration. OVERFLOW grids inthese areas, to allow for inter-grid
[1,10], aNASA research flow solver which communication through interpolation and to
uses the Chimera overset grid approach,. capture the viscous effects inthe juncture all
has been selected for this evaluation. inthe same grid. Collar grids are generated
OVERFLOW offers awide array of by identifying the intersection curve between
numerical schemes and turbulence models, components and growing a grid onto the
and has been accepted throughout the adjacent components.
aircraft industry as one of the leading Navier-
Stokes codes available. Collar grids were generated at the pylon-
nacelle and pylon-fuselage junctures. For
Grid Generation the wing-body fairing, acollar grid was
generated at the fuselage-fairing intersection
The overset grid approach has the and at the wing-fairing intersection. Athird
advantage ofallowing grid generation of grid originating at the model centerline
aircraft components separately, thus covered the remainder of the fairing.
providing gridsthat conformtothe
component topology. Through a holecutting The NASA Ames Chimera Grid Tools (CGT)
procedure andboundarypoint interpolation, software package [4] was used to generate
excess overlap between component grids is the collar grids, to add wakes to the wing
blanked out. and pylon grids, and to add awing tip cap
grid [14]. Another cap grid was required for
The geometry modeled here isa business the pylon shelf, which extended beyond the
jet configuration with aft fuselage mounted nacelle exhaust plane. To increase
pylon/nacelles. Figure 1shows the different communication with the nacelle grid, the
surface grids that make up the current pylon cap grid was extended upstream and
model. Inall fifteen grids, including the outer projected onto the nacelle surface (Figure
box grids, were needed to define the 3).
geometry.
Volume Grids
Surface Grids From the body-fitted surface grids, volume
grids were generated with HYPGEN [5-6].
The first step inthe grid generation process The initial spacing off the surface is equal to
isto generate surface grids for the various 0.00035 inches, which corresponds to ay÷
2
American Institute ofAeronautics and Astronautics
AIAA-2001-1004
value of 1at 10 percent chord from the wing create collar surface grids, shown in Figure
leading edge. 2. Second, volume grids are generated with
HYPGEN. Finally, CGT utilities are used to
Communication between the component grid add reflected symmetry planes and smooth
outer boundaries, and extension of the the wake region of the wing-fairing collar.
computational domain to the far field is
accomplished with two levels of Cartesian Flow Solver
box grids, following an approach used for
other geometries [7-8]. The OVERFLOW Navier-Stokes flow solver
is used in this analysis. This code uses an
Grid Communication implicit approximate factorization algorithm
to solve the thin-layer formulation ofthe
The PEGSUS 4.0 code [9], developed by Navier-Stokes equations. For these
CALSPAN at AEDC, is used to remove calculations, central differencing with
excess overlap between grids and find second- and fourth-order artificial dissipation
interpolation stencils for inter-grid boundary is used for the Euler terms. Trial runs with
points. This process effectively connects all Roe's upwind differencing scheme did not
of the overset component and box grids into show improvement over central differencing.
one system. Local time stepping, grid sequencing, and
multigrid are used to accelerate
Scriptin.q the Process convergence [10].
Itis recognized that this grid generation While steady-state acceleration techniques
process is iterative. Several attempts may be were used for these simulations, the mid-
made to ensure that the surface is range and high angle-of-attack cases were
adequately represented and that the number largely separated. Total liftcoefficient varied
of orphan points, grid points with no by 0.02 from a mean value in some cases.
adequate interpolation stencil, is minimized. Itis recognized that a more thorough
Also, during development phase, the analysis of the unsteady aspect of these
configuration isconstantly changing as more flowfields is needed; however, for the
refinements are introduced. With this in airloads analysis process only averaged
mind, aset of scripts that performs the tasks steady loads are desired.
highlighted above is highly desirable. This
allows for making changes to the various Processinq Requirements
components and then regenerating the
entire grid system with a minimum amount of The grid system for this configuration has a
user input. total of 3.9 million points. Solutions were
generated on an SGI Origin 200 with four
The present work uses amodified set of processors and 1.8 GB of memory. Each
scripts, originally developed for ageneric case required about 260 MB of memory and
business jet configuration. These scripts 30 hours to converge on asingle processor,
encompass surface grid generation and except for higher angles-of-attack. These
refinements, volume grid generation, hole were processed further to get the solution
cutting and interpolation stencil identification. stabilized within acertain band.
An example illustrating the script approach is
the generation of the wing-body fairing collar Results and Discussion
grids. As shown in Figure 4, this process
starts with wing and fuselage surface grids, Atotal of six flow conditions are examined in
the fairing definition inthe form of reference the current study. These consist of a series
grids (created in GridTool from the IGES of three angles-of-attack (low, mid-range
CAD definition), and grid lines representing and high) at two transonic Mach numbers
the intersection of the fairing with each of the above cruise. For each flow condition,
fuselage, wing, and symmetry plane. results were obtained using the Baldwin-
Barth [11]and Spalart-AIImaras [12] one-
The first step uses the SURGRD surface equation turbulence models, and the k-
grid generation code [15] from CGT to omega two-equation turbulence model [13].
3
American Institute of Aeronautics and Astronautics
AIAA-2001-1004
Attempts to use the Menter SST model [16] the trailing edge. Results for intermediate
were unsuccessful due to wing root grid stations (not shown here) indicate that this
issues. Surface pressure coefficients onthe pattern disappears ashort distance outboard
wing are presented at two stations, one of this station. This could bea result of
inboard and one outboard. The Reynolds solution convergence difficulty at these
number matches that at which the extreme conditions. Further investigation of
experimental data was collected inthe wind this issue is required. Although the k-omega
tunnel. model continues to predict a shock location
too far aftas inthe previous cases, the post-
Atlowangle-of-attack, where the flow shock pressure levels match wind tunnel
remains attached to the surface, solutions results. The lower surface predictions for all
using the Baldwin-Barth model show surface models remain good.
pressures invery good agreement with wind
tunnel data (Figure 5), with shock location Support for the aerodynamic loads process
and strength well predicted. Spalart-AIImaras is provided by supplying design loads for
results are similar. The k-omega model various components of the aircraft such as
predicts ashock location significantly aft of the fuselage, nacelle, pylon, etc. This is
the one-equation models. This behavior is done by providing total component load
similar to that shown in Reference 13for (e.g., pylon normal force coefficient), or by
transonic flow over the RAE 2822 airfoil. providing arunning load.
For the mid-range angle-of-attack (Figure 6), Here, running loads are computed from the
Baldwin-Barth results show shock location CFD solution by dividing the body into a
and strength are well predicted onthe upper number of segments and integrating the
surface inboard station for both Mach pressure to obtain either aforce ora
numbers and on the lower surface for the moment coefficient inthese segments. Data
higher Mach number case. The upper can bepresented either as acumulative total
surface outboard station pressure starting from aspecific point, or separately
comparison shows that the shock position is for each of the segments, as in Figure 8 for
about 5percent chord aft of that measured the fuselage. Individual component loads
inthe wind tunnel. Here the flow separates can also be used to determine the fraction of
behind the shock. For the most part the the aircraft total load that is being carried by
lower surface predictions are good. Toward aparticular component (Figure 9).
the trailing edge, the calculated pressure
coefficient shows significant unsteadiness, Ifaccurate, CFD solutions are invaluable for
varying with where the solution is stopped. the aerodynamic loads process because
This is more noticeable inthe high Mach they provide distributed surface pressures
number case. This behavior is not present in which can be analyzed component-by-
the Spalart-AIImaras results, where the flow component. These supplement wind tunnel
is better behaved at the trailing edge. The data from extensive pressure taps or
predictions on the upper surface of the component balances, both expensive
inboard station are slightly better than experimental techniques.
Baldwin-Barth. Again, the k-omega model
predicts the shock location aft of the other Conclusion
models.
A Navier-Stokes flow solver, OVERFLOW,
In the high angle-of-attack case (Figure 7), has been successfully used for the
where aconsiderable amount of separation prediction of aerodynamic loads on a
is present, the upper surface rooftop and business jet configuration. Three turbulence
shock location predicted by both one- models were evaluated at above-cruise
equation models are fairly good. However, Mach numbers and low to high angles-of-
pressures aft of the shock at the inboard attack. Overall, results show the Baldwin-
station do not match the trends ofthe wind Barth and Spalart-AIImaras models providing
tunnel data. A lower pressure atthe upper the closest match to experimental results.
surface trailing edge leads to an acceleration The k-omega turbulence model predicts the
of the flow on the lower surface approaching shock location and flow separation farther aft
4
American Institute of Aeronautics and Astronautics
AIAA-2001-1004
than the one-equation models. Further 8. R.L. Meakin, "Moving Body Overset Grid
investigation of isolated flow patterns at the Methods for Complete Aircraft Tiltrotor
middle and high angle-of-attack cases is Simulations," AIAA-93-3350, July 1993.
needed.
9. N.E. Suhs and R.W. Tramel, "PEGSUS
4.0 User's Manual," AEDC-TR-91-8, Arnold
References Engineering Development Center, Arnold
AFB, TN, Nov. 1991.
1. P.G. Buning, et al., "OVERFLOW User's
Manual, Version 1.8," NASA Langley 10. D.C. Jespersen, T.H. Pulliam, and P.G.
Research Center, Hampton, VA, Feb. 1998. Buning, "Recent Enhancements to
OVERFLOW," AIAA 97-0644, Jan. 1997.
2. J.Samareh-Abolhassani, "GridTool: A
Surface Modeling and Grid Generation 11. B.S. Baldwin and T.J. Barth, "A One-
Tool," NASA CP-3291, 1995. Equation Turbulence Transport Model for
High Reynolds Number Wall-Bounded
3. S.J. Parks, P.G. Buning, J.L. Steger, and Flows," AIAA 91-0610, Jan. 1991.
W.M. Chan, "Collar Grids for Intersecting
Geometric Components Within the Chimera 12. P.R. Spalart and S.R. AIImaras, "A One-
Overlapped Grid Scheme," AIAA 91-1587, Equation Turbulence Model for Aerodynamic
June 1991. Flows," La Recherche Aerospatiale, No. 1,
1994, pp. 5-21.
4.W.M. Chan, "Manual for Chimera Grid
Tools," NASA Ames Research Center, 13.J.E. Bardina, P.G. Huang, and T.J.
Moffett Field, CA, Oct. 1998. Coakley, 'q'urbulence Model Validation,
Testing, and Development," NASA TM
5.W.M. Chart and J.L. Steger, 110446, April 1997.
"Enhancements of aThree-Dimensional
Hyperbolic Grid Generation Scheme," Appl. 14. S.E. Rogers, H.V. Cao, and T.Y. Su,
Math. and Comput., Vol. 51, pp. 181-205, "Grid Generation For Complex High-Lift
1992. Configurations," AIAA 98-3011, June 1998.
6.W.M. Chan, I.-T. Chiu, and P.G. Buning, 15.W.M. Chan and P.G. Buning, "Surface
"User's Manual for the HYPGEN Hyperbolic Grid Generation Methods for Overset Grids,"
Grid Generator and HGUI Graphical User Computers and Fluids, Vol. 24, No. 5, 1995,
Interface," NASA TM 108791, Oct. 1993. pp. 509-522.
16. F.R. Menter, 'q-wo-Equation Eddy
7. D.G. Pearce, et al., "Development of a Viscosity Turbulence Models for Engineering
Large-Scale Chimera Grid System for the Applications," AIAA J., Vol. 32, Nov. 1994,
Space Shuttle Launch Vehicle," AIAA 93- pp. 1299-1310.
0533, Jan. 1993.
5
American Institute ofAeronautics and Astronautics
AIAA-2001-1004
Figure 1: Tail-Off Chimera Grid Model of a Business Jet
Figure 2: Wing-Body Fairing Collar Grids
6
American Institute of Aeronautics and Astronautics
AIAA-2001-1004
Figure 3: Pylon Shelf Cap Grid
Z...........
Figure 4: Winci-Bodv Fairin_qGrid Generation Process
Input (from CATIA, GridTool): Surface .qrids(from SURGRD):
Fuselage grid Fuselage-fairing collar
Wing grid Wing-fairing collar
-----b
Fairing definition Fairing-symmetry collar
Fairing intersection curves
/
Volume .qrids(from HYPGEN): Final .qrids(from CGT utilities):
Fuselage-fairing collar Fuse-fairing (with symmetry planes)
Wing-fairing collar Wing-fairing (with smoothed wake)
Fairing-symmetry collar Fairing-symmetry (with symmetry planes)
7
American Institute of Aeronautics and Astronautics
AIAA-2001-1004
Figure 5: Predicted vs. Measured Wing Pressure Coefficient: Low Angle-of-Attack
MACH=0.75 Outboard Station MACH=0.79 Outboard Station
...!...........i..............i.......J..........i..........
i
, 0
Cp ..............._...........
_ _© Wind Tu !OW;undiTun_el ! '
trth::....... ......-:..:.............i..........._-B ddwin_Earth::.......
-- ].............. [.......... _--B_tdw!n-B;
i ! -i- Si_alad.AI maras - ---- S :)ala_t-AI Ima_as
.....i............L.............._.+..i.°K...m.e,.g.a.. _ :; _-_+ k.,ome_ ga
Inboard Station Inboard Station
• ; , , ,
....... :.............. :.............. 'r...... 4....... }.............. _.......
1
! : i
_---;. ............... i............ ": " _'-'- ' ......
Cp
! _._0 Wind ,:'l'----unr
• !............ T........ -_ B_lcl_'n B;_.rth-!........ :.... _.............. _............ _- B_t_--Barth_ .......I
: i -::- S_ala_-AI Im_ (_-_ • _ :: - -i- Rbalatt-- .A!Imatas__
® _. i. _:._om_oa i ,
' ; i _-_+ k_om_)ga ....... ;.......
....... ;.................................... '
Chordwise Chordwise
Figure 6: Predicted vs. Measured Wing Pressure Coefficient: Mid-Range Angle-of-Attack
MACH=0.75 Outboard Station MACH=0.79 Outboard Station
: I i i i :
...-..i_....-..].,..._..<..-._i.._.. .....i............i..............i........
Ci ....i......?..___
Cp i-i--- "_ (z....-r_,_-
o,,,,nT,un,e; *
'- tOV_ _d'runrel i
...._ B ,ldwin-B'lrth-i,_ ....... _! .......'......._..........._B ddwin:E arthl.......
_ _ -i- S )ala_-Allma_as-
....i...........L.............i.+.ki°m_ga......i[.........
" !! i _+ k-ome_ga i
Inboard Station Inboard Station
: "\! . _. ! , :
Cp __.......i.......... _"q, .......... __.O....W....in..d.. Ti.u..n..r.,.e..l.... ii.......
...... I.......i...... I--- _-B_Id_Lm=E arth":.......
i 1 i I -!- SDala_rt-Allma(as_
¢
_m__ga .................
! i __+ kT_°m_ga :
Chordwise Chordwise
8
American Institute of Aeronautics and Astronautics
AIAA-2001-1004
Figure 7: Predicted vs. Measured Wing Pressure Coefficient: High Angle-of-Attack
MACH=0.75 Outboard Station MACH=0.79 Outboard Station
....i.-.-.T i i
: +--4-,, i i i
Cp
__eli
" i.._4v_. -'[ ioHind Tear, el ::
_ l i 1......i- Sbalaa-Al!maras-
...i..........
ii I il [_ _._+k_°me_ga i
Inboard Station Inboard Station
\! ; ;
cp ' '
" i -.,_'_, i0 V_ind ,:Tunrlel i
.......i......'---_-B _tclw)n=Earth{::.......
(_" .......i..........._B£ddw,i_nlBA:h,r-tAhi!l.m_.q ....... y _"_ iOV_'ind Tunr,el i
! -i- £ :_ala_--AI
i ,4'+ k--om,ega i
Chordwise Chordwise
Figure 8: Fuselage Normal Force Runninq Load
30 ! : !
O
.J
20
t-
t-
t-
tT"
m 10
2
O
LL
E 0
t_
o
Z
_ -10
LL
-2O
0 200 400 600
Fuselage Station (inch)
9
American Institute of Aeronautics and Astronautics
AIAA-2001-1004
Figure 9: Aircraft Components, Percent of Total Load
100
i : i ! i
o_ 80
c : i i
............. ;............. _............. :-............. _............. ;-............. t............. :-.......................... -.............
c
i i i _ O---OWing%
O
Q.
E 60 , , , ' [3- =-_ F0sela _ %
O
o :: i . i ......... i......_..: ...::F..PY!o.n._.%.
....J........I.....................[................_..-.-_.:ace,,e:
£
....i....................i.............t...........................i.....................i..................................
o 40
I!
E
p
z0 20 i : : i
i i : i
0 A , Ant le of Atta "k
10
American Institute of Aeronautics and Astronautics