Professional Documents
Culture Documents
Composite skid landing gear design feasibility results are presented. Limit drop test as per Federal Aviation Regulation
(FAR) Part 27 has been simulated by finite element analysis using ABAQUSTM for multiple landing scenarios from a drop
height of 508 mm (20 inches). Load factor, ply stresses, and ply strains are evaluation metrics. Study of seven fibers and
two matrix systems has been used for laminate selection. 48 ply all-composite designs exhibit load factors within 4–6g with
more than 40% weight saving in comparison to a 1/4-inch wall thickness aluminum design. High strains are observed in
crossmember radii. Lower load factors with weight saving increased to 49% are observed in hybridized metal–composite
configurations. Benefits of metal–composite hybridization are region dependent only. A 48-ply design with Ke49/PEEK
skids, 1/4-inch-thick Al 7075 crossmember radii, and the rest of IM7/PEEK (or IM7/977-3) shows desired performance.
Physics-based strength failure criteria are recommended over maximum strain theory.
Nomenclature of initial taxi and takeoff run capability but at the cost of design com-
plexity. Skid landing gear, on the other hand, offers design simplicity
d fuselage displacement upon impact and reduction in empty weight (WE ). Traditionally, skid landing gears
h drop height have been manufactured from elasto-plastic metal alloys, which dissipate
L ratio of assumed rotor lift to the rotorcraft weight energy during plastic bending. Federal Aviation Regulation (FAR) Part 27
n limit inertia load factor (Ref. 1) regulations permit yielding of the landing gear under limit loads.
nj load factor developed, during impact, on the mass used in the Under crash loads, the metal skids and crossmembers plastically deform,
drop test (i.e., the acceleration dv/dt in g recorded in the drop test dissipating energy while allowing for a controlled fuselage underbelly
plus 1.0) crash.
WE empty weight Reduction in gross weight (WG ) and empty weight (WE ) of the aircraft
WG gross weight are primary design concerns. Corrosion resistance concerns in metals
γ12 shear strain and fatigue performance can be adequately addressed with composite
ε11 normal strain along the longitudinal fiber direction materials. Composites offer other advantages such as high specific energy
ε22 normal strain along the transverse direction absorption (SEA) under crushing loads (Ref. 2). Literature search reveals
ε33 normal strain out of the plane that a composite skid landing gear design would be a novel idea. The
με microstrains first design step requires evaluation of a composite skid landing gear
ν12 in-plane Poisson’s ratio performance under limit loads as per FAR Part 27.725. Limit drop test
ρ density (kg/m3 ) simulation by dynamic explicit FEA was conducted by Tho et al. (Ref. 3)
σ11 normal stress along the longitudinal fiber direction using LS-DYNA and Sareen et al. (Ref. 4) using MSC/DYTRAN® .
σ12 shear stress Airoldi and Janszen (Ref. 5) have performed similar analyses using
σ22 normal stress along the transverse direction MaDyAc, a multibody tool, and optimized the metal skid landing gear
σ33 normal stress out of the plane design. This paper reports load factors and local ply stresses and strains
obtained using ABAQUSTM (Ref. 6) for all-composite, hybrid composite,
Introduction and hybrid metal–composite designs.
Rotary winged aircraft traditionally use one of two types of landing Skid Landing Gear Configuration
gear systems. The oleo-strut landing gear with wheels offers advantages
The skid landing gear was based on a Bell AH-1G Cobra helicopter.
The rear and front crossmember placements and heights were based
∗ Corresponding author: email: badger@gatech.edu. on work done by Airoldi and Janszen (Ref. 5) and Monterrubio and
Presented at the 64th American Helicopter Society Annual Forum, Montreal, Sharf (Ref. 7) as complete data were unavailable. A linear taper law
Canada, April 29–May 1, 2008. Manuscript received December 2007; accepted was incorporated in both crossmember horizontal beam regions. Figure
July 2009. 1 shows a wireframe model. The rear crossmember is 490 mm in height,
Fig. 1. Skid landing gear configuration. E 1 (GPa) E 2 (GPa) G12 (GPa) ν12 ρ (kg/m3 )
Fiber
IM7 292 19.5 70 0.2 1790
80 mm lower than the front crossmember, and the skid track/tread is 2140 AS4 231 14 14 0.2 1790
mm. The center of gravity (CG) location is 770 mm along the station T700S 230 23.2a 95.8b 0.2 1800
line (STA), from the rear crossmember, 0 mm from the butt line (BL), Ke49 130 13c 13d 0.2 1450
and 1854.2 mm from the skids along the water line (WL). Helicopter Matrix
specifications are shown in Table 1 as taken from Ref. 8. Composite skid 977-3 3.45 3.45 1.23 0.32 1300
PEEK 4 4 1.42 0.4 1309
inner diameters and crossmembers have been sized for a typical 1/4-inch
Metal
wall thickness. Outer diameters of the skid tubes and crossmembers are Al 7075 72 72 27.03 0.33 2700
101.6 mm (4 inches) and 76.2 mm (3 inches) with maximum taper at the
a−d
crossmember midspan, resulting in a diameter of 101.6 mm (4 inches). Approximated.
Federal Aviation Regulation (FAR) Part 27.725 eliminated due to high densities, which hinder weight saving. All fibers
are transversely isotropic, and matrices are isotropic. Table 3 lists the
FAR Part 27.725 requires the drop height to be no less than 203.2 mm fiber/matrix properties (Refs. 9–13). Transverse shear stiffness (G23 ) for
(8 inches) and as generally 330.2 mm (13 inches). A typical 508 mm IM7 was used as 5.74 GPa. Elastic properties for unidirectional lam-
(20 inches) drop height for military helicopters was chosen. Load factors ina have been calculated in Matlab© V 7.0 (Ref. 14) using the rule of
were calculated using Eqs. (1) and (2). The ratio of rotor lift to the mixtures. Table 4 lists comparison of IM7/977-3 lamina properties with
rotorcraft weight was prescribed a value of 0.67. experimental results for a 65% fiber volume fraction laminate published
h + (1 − L) d by Kulkarni (Ref. 15). Results match within 13% for all moduli values.
We = W × (1) Table 5 lists composite lamina properties. An N = 4 (π /4) 48-ply quasi-
h+d
isotropic [(45/–45/0/90)6 ]s laminate with ply thickness of 0.127 mm
We
n = nj × +L (2) (5 mils) has been used. Equivalent laminate properties have been com-
W
puted in Matlab© V 7.0 as listed in Table 6. To eliminate testing, a linear
spring-mass model has been used for load factor comparison. The mate-
Material Selection rial with the lowest stiffness has the lowest load factor. Static analysis has
been done for a lower bound on stiffness based on strains. Ke49 was elim-
Al 7075 has been used for the reference landing gear as comparison. inated due to low stiffness. Finite element analysis (FEA) has been done
Seven unidirectional tape fibers and two matrix systems were consid- for remaining fibers with thermoplastic poly-ether-ether-ketone (PEEK)
ered for comparison of load factors and strain limits. Fiber properties and 350◦ F cure 977-3 epoxy resin as the matrices.
(Refs. 9–12) are shown in Table 2. S-Glass, E-Glass, and boron were
Finite Element Analysis
Table 1. Helicopter design data Dynamic explicit FEA was conducted using ABAQUS/EXPLICIT V
Design Parameter Metric Value 6.7-1. Skid tubes and crossmembers were modeled with conventional
042004-2
COMPOSITE SKID LANDING GEAR DESIGN FEASIBILITY 2009
Table 5. Lamina properties (65% Vf ) Table 7. Typical laminate strain allowable (103 με)
Laminate E 1 (GPa) E 2 (GPa) G12 (GPa) ν12 ρ (kg/m3 ) Fiber Angle ε11 ε22 ε33
IM7/977-3 191 9.5 5.3 0.24 1622 0◦ 14 8 15
IM7/PEEK 191.2 10.7 5.76 0.25 1619 90◦ 10 8 15
AS4/977-3 151.4 8.2 4.17 0.23 1622 ±45◦ 6 8 15
AS4/PEEK 151.6 9.1 4.45 0.26 1619
T700S/977-3 150.7 10.2 5.40 0.23 1628
T700S/PEEK 150.9 11.1 6.21 0.23 1625
Ke49/977-3 85.7 7.9 4.09 0.23 1401 Table 8. Typical laminate stress (MPa) allowable
Ke49/PEEK 85.9 8.7 4.36 0.26 1398 (tensile/compressive)
042004-3
K. SHROTRI JOURNAL OF THE AMERICAN HELICOPTER SOCIETY
Level landing: Composite skid landing gears IM7 fiber shows better performance in the crossmember regions for a
single all-composite configuration. It is to be noted though that the skids
Figure 5 shows CG displacements and accelerations of the 48-ply all- produce acceptable strains. Hence, even low stiffness yet high impact
composite IM7/977-3, T700S/977-3, AS4/977-3, and IM7/PEEK skid resistant Ke49 fiber may be used in this region. A key point to be noted is
landing gears. As expected, accelerations are higher than for the metal the absence of plasticity and nonlinearity in the stress–strain constitutive
landing gear. The highest load factors computed are 5.75g and 5.73g law prescribed for the composites. Any nonlinearity and possible plas-
for IM7/PEEK and IM7/977-3 designs, respectively. Figures 6 and 7 ticity effects displayed by the composite will result in energy dissipation,
show that normal and shear stresses are greater than 3000 and 124 MPa, hence lower load factors and also more acceptable strains and stresses.
in the 0◦ and 45◦ fibers, respectively, and exceed allowable values. Hybridization of composites and metal–composites was also analyzed to
Figures 8–15 show normal strains along the 0◦ fiber in ply 46 and shear study the effect on load factor and strain/stress relief.
strains along the 45◦ fiber in ply 48 along the crossmember outer diam-
eters. Most of the IM7 fiber crossmember beam produce strains within Level landing: Hybrid skid landing gears
±3000–6000 με. Ply strains are high in localized regions close to ra-
dius bends on the rear crossmember, indicating that local stiffening or From previous results, it can be noted that strains are large in re-
plasticity induced strain relief may be required. Shear strains are as high gions close to the bend radii of the crossmembers. Strains produced by
as ±10,000–30,000 με in the rear crossmember. Both strains increase the Ke49 skids are within acceptable values. Ke49 fiber can be used
with decreasing fiber stiffness. Thus, it can be said that high stiffness for the skids leaving the remainder with high stiffness fibers. To allevi-
ate strains in the crossmembers, the radii bends were assigned Al 7075
elasto-plastic properties. Thus, a combination of metal plasticity with
the nonlinear constitutive stress–strain law and linear composite stress–
strain law was prescribed. Several configurations, with both PEEK and
977-3, were analyzed to understand the amount of strain relief as well as
the volume fraction dependence of metal in the design. Figure 16 shows
the CG displacements and accelerations of the hybrid designs. It can be
seen that metal plasticity in the bend regions alleviates the load factors
significantly without adding much weight. The lowest load factor with
acceptable strains is obtained for a design with Ke49/PEEK skids, Al
7075 crossmember bends, and the rest IM7/977-3. As the elastic prop-
erties of PEEK are very nearly similar to those of 977-3, it is projected
that replacing IM7/977-3 with IM7/PEEK in any configuration should
result in near identical performance. Figures 17–36 show normal strains,
normal stresses, shear strains, and shear stresses in the 0◦ and 45◦ fibers
in plies 46 and 48 along the outer diameters of the rear crossmember.
Fig. 4. Von Mises stress. Figure 37 shows out-of-plane strains. IM7 produces 7000 με in the
042004-4
COMPOSITE SKID LANDING GEAR DESIGN FEASIBILITY 2009
0◦ fiber and 9000–12,000 με along ±45◦ in shear. AS4 and T700S pro-
duce marginally higher strains. Both, normal and shear strains are high.
Strength-based theory shows that any of the three fibers is acceptable
provided the regions close to the crossmember bend radii are metallic.
While IM7 fiber shows the most desired performance, stress-based al-
lowable show that any of the fibers can be used in conjunction with
metal plasticity in the bend regions. Hence, AS4 and IM7 fiber, with
high SEA values (Ref. 17), are recommended. It can be concluded that
metal plasticity aids significantly in lowering the load factor. Thus, it
may be possible to use either of the fibers in conjunction with 977-3,
PEEK, or a similar matrix, for the hybrid metal composite design. The
configuration with Al crossmember bends and skids and the rest made
of IM7/977-3 has a load factor of 4.40g while that with only Al bends Fig. 8. Ply 46 (0◦ ) ε11 : IM7/977-3.
042004-5
K. SHROTRI JOURNAL OF THE AMERICAN HELICOPTER SOCIETY
042004-6
COMPOSITE SKID LANDING GEAR DESIGN FEASIBILITY 2009
042004-7
K. SHROTRI JOURNAL OF THE AMERICAN HELICOPTER SOCIETY
042004-8
COMPOSITE SKID LANDING GEAR DESIGN FEASIBILITY 2009
Fig. 25. Ply 46 (0◦ ) ε11 : Ke49 skids/IM7/PEEK/Al bends. Fig. 28. Ply 48 (45◦ ) σ12 : Ke49 skids/IM7/PEEK/Al bends.
Fig. 27. Ply 46 (0◦ ) σ11 : Ke49 skids/IM7/PEEK/Al bends. Fig. 30. Ply 48 (45◦ ) γ12 : Ke49 skids/AS4/PEEK/Al bends.
042004-9
K. SHROTRI JOURNAL OF THE AMERICAN HELICOPTER SOCIETY
Fig. 31. Ply 46 (0◦ ) σ11 : Ke49 skids/AS4/PEEK/Al bends. Fig. 34. Ply 48 (45◦ ) γ12 : Ke49 skids/T700S/PEEK/Al bends.
Fig. 33. Ply 46 (0◦ ) ε11 : Ke49 skids/T700S/PEEK/Al bends. Fig. 36. Ply 48 (45◦ ) σ12 : Ke49 skids/T700S/PEEK/Al bends.
042004-10
COMPOSITE SKID LANDING GEAR DESIGN FEASIBILITY 2009
Fig. 38. Level landing with drag: CG vertical displacements and accelerations.
042004-11
K. SHROTRI JOURNAL OF THE AMERICAN HELICOPTER SOCIETY
Fig. 41. Ply 46 (0◦ ) ε11 : IM7/977-3 for rolled attitude landing.
Fig. 39. Ply 46 (0◦ ) ε11 : IM7/977-3 for level landing with drag. should result in higher computational expense. Thus, efficient modeling
techniques for dynamic analysis of composites are needed. For validation
above or close to typical allowable for such fibers in the all-composite of results, the reference aluminum skid landing gear had been analyzed
designs. Metal–composite hybridization offers better designs for any of to ensure that the resulting displacements, reactions and hence the load
the three fibers. Allowable values used are typical values based on manu- factors computed are within the correct range.
facturer data sheets and published data, which in fact report higher stress Metal–composite hybridization takes advantage of plasticity, thus re-
allowable for individual fibers. Thus, the designs may be feasible based sulting in energy dissipation and lower load factors. From the crashwor-
on strength theory than on maximum strain theory. thy perspective, it may be useful to have plastic energy dissipation in these
It is to be noted that hourglass effects could be of concern in the regions, as the behavior of composite tubes under impact when at an in-
analyses. Coarse mesh and reduced integration elements are believed to clination may not produce crushing (Ref. 19) and hence, increased SEA.
be the cause of this. Hourglass control with enhancement can be used in Fabricating the radii bends from elasto-plastic metal and crossmember
addition with mesh refinement. However, mesh refinement will mean an from acceptably low stiffness fiber, such as AS4, to increase the bending
expensive run time when analyzing composites. While full integration moment versus rotation area under the curve, may be useful. Only the
elements such as the S4 element or higher order elements such as the eight horizontal tapered crossmember beams under bending loads and the skid
node reduced integration shell element (S8R) may be better, these too tubes under compressive and shear loads would be the composite region
042004-12
COMPOSITE SKID LANDING GEAR DESIGN FEASIBILITY 2009
042004-13