Академический Документы
Профессиональный Документы
Культура Документы
Vibratory loads on helicopters arise from a variety The elemental lift and drag forces can be written
of sources such as the main rotor system, the from simple strip theory. The blade is seperated into
aerodynamic interaction between the rotor and the several stations and the cross-section of the blade at
fuselage, the tail rotor, the engine and the one of the stations and the velocity components,
transmission. However, the most significant source forces and angles at this cross-section are shown in
of vibration in a helicopter is the main rotor because Fig.1.
of the unsteady aerodynamic environment acting on
highly flexible rotating blades. The reduction of Fz
vibration levels in helicopters below acceptable limits L
is one of the main problems facing rotorcraft
designers. Vibrations lead to passenger discomfort,
pilot fatigue, increased noise levels, degradation of
weapon effectiveness, and premature failure of
aircraft parts. Fy
The field of helicopter aeroelasticity has been a U D
very active area of research during the last 40 years.
A comprehensive review paper about this area has UP
θ α
been published recently [5]. φ
In this study the blade is modeled as a slender, UT
deformable beam composed of isotropic and
homogeneous material. The Euler-Bernoulli Figure 1. Velocity components, forces and angles at
hypothesis is assumed to apply and the Blade the blade cross section
Element Theory is used to obtain aerodynamic
loadS. A semi analytical-numerical technique
1 2
Assosciate Professor, E-mail: kayam@itu.edu.tr Research Assistant, E-mail: ozdemirozg@itu.edu.tr
AIAC-2005-054 Kaya & Özdemir
ρac
Here L is the lift force, D is the drag force, Fz
dF y = −
2
[θU P ]
U T + U T2 C d / a − U P2 dx (7)
is the vertical force, Fy is the inplane force. U p ,
U T and U are the perpendicular velocity Incorporating the airfoil flap and lead
component, tangential velocity component and displacements (v,w) in the rotating coordinate
system, the relative velocities become
resultant velocity, respectively. θ , ϕ and α
represent, the geomentric pitch angle, the induced U T = Ωx + v& (8)
inflow angle and the effective angle of attack,
respectively. U P = υ i + w& + Ωvβ pc − (v& + Ωx )(θ + φ ) (9)
Some simplifying assuptions used in the Substituting Eqs. (8) and (9) into the force
formulation of aerodynamic forces are as follows: equations and discarding second order products of
2 2
1. Since the tangential velocity component is much displacement velocities such as v& , w
& , v&w& , the
greater than the vertical velocity component final equations are obtained.
ρac 2 2 CD
U T >> U p and U = U P +U T ≈ U T − Ωxν i (θ + φ ) −
2 2
Fy = νi − Ω x
2
2 a
(10)
2. Since inflow angle is small, CD
tan ϕ ≅ ϕ = U P U T . Therefore, the effective angle 2Ωx a + ν i (θ + φ )v& + [2ν i − Ωx(θ + φ )]w&
of attack can be written as
3.
ρac
α = θ −ϕ ≅ θ −
Up
(1)
Fz =
2
{− Ωxν i + Ω 2 x 2 (θ + φ ) −
UT
xc
Ω 2 xβ pc v + Ω 2(β pc + w′) − Ωxw& + (11)
The expressions for the sectional lift and the 2
sectional drag forces are given by 3 c
cΩxφ& + [2Ωx(θ + φ ) −ν i ]v& − w
&&
1 1 4 4
L= ρU 2 cCl = ρU T2 caα (2)
2 2
ρac c 2
Mφ = − Ωxφ& (12)
1 1 C 2 8
D = ρU 2 cC d = ρU T2 ca d (3)
2 2 a
Here, M φ is the aerodynamic moment, β pc is the
Here; ρ is tha air density, a is the lift curve slope, precone angle, Ω is the constant rotation speed, φ
c is the chord length, Cl is the sectional lift is the torsion angle and vi is the induced inflow
coefficient and C d is the sectional drag coefficient. velocity.
Structural Modelling
Considering the assumptions and Fig.1, the
vertical and the inplane forces can be written as In this section, Hamilton’s principle is used to
follows: derive the partial differential equations of motion. In
Fig.2, rotor blade cross-section before and after
dFz = dLCosϕ − dDSinϕ ≅ dL − dDϕ (4) deformation is shown.
ρ ac
dF z =
2
[θU 2
T ]
− (1 + C d / a )U pU T dx (6)
2
Ankara International Aerospace Conference
AIAC-2005-054 Kaya & Özdemir
Here, k A is the blade cross section mass radius
of gyration, k m1 and k m 2 are the principal mass radii
of gyration, m is the mass per unit length and T is
the centrifugal force that varies along the spanwise
direction and which is expressed as follows
L
L2 − x2
T ( x) = ∫ ρAΩ2 (r + x)dx = ρAΩ2 r(L − x) + (17)
x 2
′
[
− (Tv ′) + EI z − (EI z − EI y )Sin 2θ v "" + ] v ( x, t ) = v 0 ( x ) + v ( x, t ) (20)
(EI − EI ) Sin2θ w
z y
""
− 2mΩβ pc w& + (14) w( x, t ) = w0 ( x) + w ( x, t ) (21)
2
m(v&& − Ω v ) = F2
y φ ( x, t ) = φ 0 ( x ) + φ ( x , t ) (22)
[
− (Tw′) + EI y + (EI z − EI y )Sin 2θ w"" +
′
]
[
− (Tv ′) + EI z − (EI z − EI y )Sin 2θ v "" +
′
]
(EI z − EI y ) Sin2θ v "" + 2mΩβ pc v& + (15) Sin 2θ ""
2 (EI z − EI y )
w − 2mΩβ pc w& +
&& = Fz − mΩ 2 β pc x 2
mw
ρac
(
m v&& − Ω 2 v = ) − Ωxν iφ −
2
(23)
δφ :
CD & &
′ ′ 2Ωx a + ν iθ v + [2ν i − Ωxθ ]w
− (GJφ ′) − k A2 (Tφ ′) + mk m2 φ&& +
( )
mΩ 2 k m2 2 − k m21 φCos 2θ = M φ − (16)
mΩ 2 (k 2
m2 − k m21 ) Sin22θ
3
Ankara International Aerospace Conference
AIAC-2005-054 Kaya & Özdemir
[
− (Tw ′) + EI y + (EI z − EI y )Sin 2θ w "" +
′
]
d dφ 2 d dφ
Sin 2θ ""
(EI − EI y )
− GJ − k A T − ω mkmφ +
2 2
(25) x r 3ρacR ν
ρac c 2 ξ= ,δ = ,γ = , λi = i ,
mΩ k 2
( 2
m2 −k 2
m1 )
φ Cos2θ = − Ωxφ&
2 8
R R m ΩR
c ω V W
c= ,ω = , V (ξ ) = , W (ξ ) = (32)
R R R R
For the flutter case, a sinusoidal variation of
v( x, t ) , w( x, t ) and φ ( x, t ) with a circular natural The centrifugal force, T , can be expressed in the
frequency ω is assumed and the functions are nondimensional form by using the first two
approximated as dimensionless parameters
v ( x, t ) = V ( x )e iωt 1 − ξ 2
(26) T (ξ ) = ρAΩ 2 + δ (1 − ξ ) (33)
2
w (x, t ) = W (x )e iωt (27) Using these parameters and Eq.(33), the Eqs.(29)-
(31) can be given by
φ ( x, t ) = φ (x )e iωt (28)
Sin2θ d 4W A7 ξW = 0
(EIz − EI y ) − 2iωΩmβ pcW − ω 2 mV −
(29)
2 dx4
ρac CD
Ω2 mV = − Ωxν iφ − iω 2Ωx +ν iθ V d 1 − ξ 2 dW d 4W
2 a − + δ (1 − ξ ) +
1 B +
dξ 2 dξ dξ 4
+ iω[2ν i − Ωxθ ]W }
d 4V
B2 + B3V + B 4W + B5ξ 2φ + B6 ξV + (35)
dξ 4
−
d dW
T [
+ EI y + (EI z − EI y )Sin θ
2
]
d 4W
+ B7 ξW + B9 ξ
dW
+ B9 ξφ = 0
dx dx dx 4 dξ
(EI z − EI y ) Sin2θ d V4 + 2iωΩmβ pcV − ω 2 mW
4
2 dx (30)
d 1 − ξ 2 dφ
ρac 2 2 xc 2 dW − C1 + δ (1 − ξ ) +
= Ω x φ − Ω β pc xV + Ω −
2
2 2 dx dξ 2 dξ (36)
3 c d 2φ
iωΩxW + iω cΩxφ + iω(2Ωxθ − vi )V + ω 2 W C2 + C 3φ + C 4ξφ = 0
4 4 dξ 2
4
Ankara International Aerospace Conference
AIAC-2005-054 Kaya & Özdemir
where the dimensionless coefficients are THE DIFFERENTIAL TRANSFORM METHOD
γ cγ cγ []
where f ( x ) is the original function and F k is the
B 7 = iω , B8 = − , B9 = −iω
6 12 8 transformed function. The inverse transformation is
2 defined as
k GJ c 2γ
C1 = A , C2 = − , C 4 = iω ∞
R mΩ 2 R 4 48
f ( x) = ∑ ( x − x0 ) k F [k ] (40)
k =0
k k k
2 2 2
C3 = −ω 2 m + m 2 − m1 Cos 2θ
R R R Combining Eqs. (42) and (43), we get
∞
( x − x0 ) k d k f ( x)
f ( x) = ∑ k
(41)
The dimensionless boundary conditions are k =0 k! dx x = x0
at ξ = 0, V =W =φ = 0 (37)
Considering Eq.(44), it is noticed that the concept
dV dW
= =0 of differential transform is derived from Taylor series
dξ dξ expansion. However, the method does not evaluate
the derivatives symbolically.
dφ
at ξ = 1, =0 (38) In actual applications, the function f ( x ) is
dξ
expressed by a finite series and Eq. (44) can be
d 2V d 2W written as follows
= =0
dξ 2 dξ 2
m
( x − x0 ) k d k f ( x)
d 3V d 3W f ( x) = ∑ (42)
= =0 k! k
dx x = x0
dξ 3 dξ 3 k =0
f ( x) = ∑ k! k
is negligibly
dx
(1) = 0 ∑ kF (k ) = 0
k = m +1 dx x = x0 k =0
df (k + 1)(k + 2 ) δ +
1
W [k + 2 ] +
(0) = 0 F (0) = 0 2
dx
δ (k + 1) W [k + 1] +
2
(44)
k (k + 1)
df
(0) = 0 F (1) = 0 2 + B 4 + kB 8 W [k ] +
dx
x=0 B 7 W [k − 1] + B 3V [k ] + B 6V [k − 1] +
d2 f
(0) = 0 F (2) = 0 B 9 φ [k − 1] + B 5 φ [k − 2 ] = 0
dx 2
1
d3 f C 2 − C 1 δ + 2 (k + 1)(k + 2 )φ [k + 2 ] +
(0) = 0 F ( 3) = 0
dx 3
C 1 (k + 1) δφ [k + 1] + C 4 φ [k − 1] +
2
(45)
∞
k (k + 1)
f (1) = 0 ∑ F (k ) = 0 C 3 + C 1 φ [k ] = 0
x =1 k =0 2
6
Ankara International Aerospace Conference
AIAC-2005-054 Kaya & Özdemir
Considering the theorems in Table 2 and 2.00
φ [0] = 0 , φ [1] = φ1
where v 2 , v3 , w2 , w3 and φ1 are arbitrary 1.40
constants.
1.20
RESULTS AND DISCUSSIONS
The computer package Mathematica is used to
write a computer program for the expressions
obtained using DTM. The results are compared with 1.00
the ones in Refs.[10]-[11] and a very good 0.00 0.40 0.80 1.20 1.60
agreement with the figures in these references is Nonrotating Fundamental Flap Frequencies
obtained.
Figure 3(a) Rotating natural flap frequency versus
When the aerodynamic terms in the rotor blade nonrotating flap frequency ( θ = 0 )
equations are eliminated, the resulting perturbation
solutions provide the rotating natural frequencies of 1.60
the rotor blade motion. For untwisted blades at zero
pitch angle (θ = 0 ) , the equations of motion are
Rotating Fundamental Lag Frequencies
In Fig.4(a)-(b), the first two flap and first two lead- 0.80
lag uncoupled rotating frequencies are given. In
Fig.4(a), values for a soft inplane rotor and in
Fig.4(b), values for a stiff inplane rotor are
introduced. For the soft inplane rotors, the first
rotating lead-lag frequency is less than the nominal 0.40
rotational frequency of the rotor while for the stiff
inplane rotors, the lead-lag frequency is greater than
the nominal rotational frequency. In this figure, Ω 0
is the nominal angular velocity and ω NR is the 0.00
natural frequency that is made dimensionless with 0.00 0.40 0.80 1.20 1.60
respect to the nominal angular velocity. Nonrotating Fundamental Lag Frequencies
7
Ankara International Aerospace Conference
AIAC-2005-054 Kaya & Özdemir
12.00 10.00
8.00
8.00
6.00
ω
Ω0 2nd Flap
4.00
4.00
1st Flap
0.00
0.00 4.00 8.00 12.00 0.00 0.20 Ω0.40 0.60 0.80 1.00
Nonrotating Fundamental Torsion Frequencies Ω0
Figure 4(b). Variation of the first two uncoupled
Figure 3(c) Rotating natural torsion frequency natural frequencies with respect to the rotor angular
versus nonrotating torsion frequency ( θ = 0 , velocity ( ω βNR = 0.6 , ωξNR = 1.5 ).
(k A k m ) = 1 .5 , k A k m = 0 )
2
8
Ankara International Aerospace Conference
AIAC-2005-054 Kaya & Özdemir
1.60 [8] Chiou JS and Tzeng JR, Application of The
Taylor Transform to Nonlinear Vibration Problems,
ωξ NR = 1.4 Transaction of the American Society of Mechanical
Engineers, Journal of Vibration and Acoustics, 118,
1st Flap Mode
Imaginary Component
p: 83–87, 1996.
[9] Zhou JK, Differential Transformation and Its
Application for Electrical Circuits (in Chinese).
0 .6 ωξ NR = 1.2 China: Huazhong University Press, P. R., 1986.
0 .8 1.20 [10] Dewey H.H. and Ormiston R.A., Stability of
Elastic Bending and Torsion of Uniform Cantilever
1st Lag Mode Rotor Blades in Hover with Variable Structural
Coupling, NASA Technical Note, 1976
[11] Dewey H.H. and Ormiston R.A., Nonlinear
ωξ NR = 1.4 1 .2 ωξ NR = 0.8 Equations for Bending of Rotating Beams with
Application to linear Flap-Lag Stability of Hingeless
Rotors, 1973.
0.80
ωξ NR = 0.6
References
9
Ankara International Aerospace Conference