Linearized Longitudinal Robert Stengel, Aircraft Flight Dynamics • 6th-order -> 4th-order -> hybrid equations" • Dynamic stability derivatives " • Phugoid mode" • Short-period mode" Cop
Trang 1Linearized Longitudinal
Robert Stengel, Aircraft Flight Dynamics
• 6th-order -> 4th-order -> hybrid equations"
• Dynamic stability derivatives "
• Phugoid mode"
• Short-period mode"
Copyright 2012 by Robert Stengel All rights reserved For educational use only.!
http://www.princeton.edu/~stengel/MAE331.html ! http://www.princeton.edu/~stengel/FlightDynamics.html !
• Symmetric aircraft"
• Motions in the vertical plane"
• Flat earth "
x1
x2
x3
x4
x5
x6
!
"
#
#
#
#
#
#
#
#
$
%
&
&
&
&
&
&
&
&
= xLon6=
u w x z q
θ
!
"
#
#
#
#
#
#
#
$
%
&
&
&
&
&
&
&
=
Axial Velocity Vertical Velocity Range Altitude(–) Pitch Rate Pitch Angle
!
"
#
#
#
#
#
#
#
#
$
%
&
&
&
&
&
&
&
&
u = X / m − gsinθ − qw
w = Z / m + g cosθ + qu
x I= cosθ( )u + sinθ( )w
z I= − sinθ( )u + cosθ( )w
q = M / I yy
θ = q
State Vector, 6 components!
Nonlinear Dynamic Equations!
Fairchild-Republic A-10!
u = f1= X / m − g sin θ − qw
w = f2 = Z / m + g cos θ + qu
q = f3 = M / Iyy
x1
x2
x3
x4
!
"
#
#
#
#
#
$
%
&
&
&
&
&
= xLon4 =
u w q
θ
!
"
#
#
#
#
$
%
&
&
&
&
=
Axial Velocity, m/s Vertical Velocity, m/s Pitch Rate, rad/s Pitch Angle, rad
!
"
#
#
#
#
#
$
%
&
&
&
&
&
State Vector, 4 components!
Fourth-Order Hybrid Equations of Motion
Trang 2Transform Longitudinal Velocity Components"
u = f1= X / m − g sinθ− qw
w = f2= Z / m + g cosθ+ qu
q = f3= M / I yy
θ= f4 = q
V = f1= T cos$% (α + i)− D − mgsinγ &' m
γ = f2= T sin$% (α + i)+ L − mg cosγ &' mV
q = f3= M / I yy
θ = f4= q x1
x2
x3
x4
!
"
#
#
#
$
%
&
&
&
=
u
w
q
θ
!
"
#
#
$
%
&
&=
Axial Velocity Vertical Velocity Pitch Rate Pitch Angle
!
"
#
#
#
$
%
&
&
&
x1 x2 x3 x4
!
"
#
#
#
$
%
&
&
&
=
V
γ
q
θ
!
"
#
#
#
$
%
&
&
&
=
Velocity Flight Path Angle Pitch Rate Pitch Angle
!
"
#
#
#
$
%
&
&
&
i = Incidence angle of the thrust vector with respect to the centerline
• Replace X and Z by T, D, and L"
Hybrid Longitudinal Equations of Motion"
V = f1= T cos$% (α + i)− D − mgsinγ &' m
γ = f2= T sin$% (α + i)+ L − mg cosγ &' mV
q = f3= M / I yy
θ = f4= q
V = f1= T cos$% (α + i)− D − mgsinγ &' m
γ = f2= T sin$% (α + i)+ L − mg cosγ &' mV
q = f3= M / I yy
α = f4 = θ − γ = q − 1
mV$%T sin α + i( )+ L − mg cosγ&'
x1
x2
x3
x4
!
"
#
#
#
$
%
&
&
&
=
V
γ
q
θ
!
"
#
#
#
$
%
&
&
&
=
Velocity Flight Path Angle Pitch Rate Pitch Angle
!
"
#
#
#
$
%
&
&
&
x1
x2
x3
x4
!
"
#
#
#
$
%
&
&
&
=
V
γ
q
α
!
"
#
#
#
$
%
&
&
&
=
Velocity Flight Path Angle Pitch Rate Angle of Attack
!
"
#
#
#
$
%
&
&
&
• Replace pitch angle by angle of attack! α = θ − γ
θ = α +γ
Why Transform Equations and
State Vector?"
• Phugoid (long-period) mode is primarily
described by velocity and flight path angle"
• Short-period mode is primarily described
by pitch rate and angle of attack"
x1
x2
x3
x4
!
"
#
#
#
#
#
$
%
&
&
&
&
&
=
V
γ
q
α
!
"
#
#
#
#
$
%
&
&
&
&
=
Velocity Flight Path Angle Pitch Rate Angle of Attack
!
"
#
#
#
#
#
$
%
&
&
&
&
&
Why Transform Equations and State Vector?"
• Hybrid linearized equations allow the two modes to be examined separately"
FLon = FPh FSP
Ph
FPh SP FSP
!
"
#
#
$
%
&
&
Effects of phugoid perturbations on phugoid motion"
Effects of phugoid perturbations on short-period motion"
Effects of short-period perturbations
on phugoid motion"
Effects of short-period period motion"
= FPh small
small FSP
!
"
#
#
$
%
&
& ≈
FPh 0
0 FSP
!
"
#
#
$
%
&
&
Trang 3Nominal Equations of Motion in
VN= 0 = f1= T cos $% ( αN+ i ) − D − mgsinγN&' m
γN = 0 = f2= T sin $% ( αN+ i ) + L − mg cosγN&' mV
qN= 0 = f3= M Iyy
αN= 0 = f4= q − 1
mV $% T sin α ( N+ i ) + L − mg cosγN&'
xN
T
=# V N γN 0 αN
$ %&
T
Equations of Motion
Sensitivity Matrices for
Longitudinal LTI Model"
ΔxLon(t) = FLonΔxLon(t) + GLonΔuLon(t) + LLonΔwLon(t)
F =
∂ f1
∂V
∂ f1
∂γ
∂ f1
∂q
∂ f1
∂α
∂ f2
∂V
∂ f2
∂γ
∂ f2
∂q
∂ f2
∂α
∂ f3
∂V
∂ f3
∂γ
∂ f3
∂q
∂ f3
∂α
∂ f4
∂V
∂ f4
∂γ
∂ f4
∂q
∂ f4
∂α
$
%
&
&
&
&
&
&
&
&
&
&
&
'
(
) ) ) ) ) ) ) ) ) ) )
G =
∂ f1
∂δ E
∂ f1
∂δT
∂ f1
∂δ F
∂ f2
∂δ E
∂ f2
∂δT
∂ f2
∂δ F
∂ f3
∂δ E
∂ f3
∂δT
∂ f3
∂δ F
∂ f4
∂δ E
∂ f4
∂δT
∂ f4
∂δ F
#
$
%
%
%
%
%
%
%
%
%
%
&
'
( ( ( ( ( ( ( ( ( (
L =
∂ f1
∂V wind
∂ f1
∂αwind
∂ f2
∂V wind
∂ f2
∂αwind
∂ f3
∂V wind
∂ f3
∂αwind
∂ f4
∂V wind
∂ f4
∂αwind
#
$
%
%
%
%
%
%
%
%
%
%
%
&
'
( ( ( ( ( ( ( ( ( ( (
Velocity Dynamics"
V = f1 =1
m[T cosα − D − mgsinγ]= 1
m C TcosαρV
2
2 S − C D
ρV2
2 S − mgsinγ
%
&
( )
• Nonlinear equation"
Thrust incidence angle neglected!
• First row of linearized dynamic equation"
Δ V (t) = ∂ f1
∂ V ΔV (t)+
∂ f1
∂γ Δ γ (t)+
∂ f1
∂ q Δq(t)+
∂ f1
∂α Δ (t)
%
&
)
*
+ ∂ f1
∂δ E Δ δ E(t)+
∂ f1
∂δ T Δ δ T (t)+
∂ f1
∂δ F Δ δ F(t)
%
&'
( )*
+ ∂ f1
%
&
)
*
Trang 4∂ f1
1
m ( CT V cosαN − CD V) ρVN
2
2 S + C ( T N cosαN − CD N) ρVNS
%
&
)
*
∂ f1
−1
m [ mg cosγN] = −g cosγN
∂ f1
−1
m CD q
ρVN2
%
&
)
*
∂ f1
−1
m ( CT NsinαN + CDα) ρVN
2
%
&
)
*
• Coefficients in first row of F"
Sensitivity of Velocity Dynamics
to State Perturbations "
C TV≡∂C T
∂V
C DV≡∂C D
∂V
C Dq≡∂C D
∂ q
C D
α ≡∂C D
∂α
V = C( Tcosα− C D)ρV
2
2 S − mgsinγ
%
&
( )
* m
Sensitivity of Velocity Dynamics to Control and Disturbance Perturbations "
∂ f1
−1
m CD δ E
ρ VN2
%
&
)
*
∂ f1
1
m CT δT cos αN
ρ VN2
%
&
)
*
∂ f1
−1
m CD δ F
ρ VN2
%
&
)
*
• Coefficients in first rows of G and L"
∂ f1
∂ f1
∂V
∂ f1
∂ f1
∂α
C T δT ≡∂C T
∂δT
C D δ E≡∂C D
∂δ E
C D δ F≡∂C D
∂δ F
∂f2
∂V =
1
mV N (C T VsinαN+C L V)ρV N
2
2 S + C( T NsinαN + C L N)ρV N S
$
%
' (
− 1
mV N2 (C T NsinαN + C L N)ρV N
2
2 S − mg cosγN
$
%
( )
∂f2
∂γ =
1
mV N[mg sinγN]= g sinγ N V N
∂f2
∂q=
1
mV N C L q
ρV N2
2 S
$
%
' (
∂f2
∂α =
1
mV N (C T NcosαN + C Lα)ρV N
2
2 S
$
%
( )
• Coefficients in second row of F"
Sensitivity of Flight Path Angle
Dynamics to State Perturbations "
• Coefficients in second row of G and L in Supplemental Slide!
C T V≡ C T
∂V
C L V≡ C L
∂V
C L q≡ C L
∂
C Lα ≡ C L
∂α
γ= C( Tsinα+ C L)ρ
2
2 S − mg cosγ
%
&
)
* mV
∂ f3
1
Iyy Cm V
ρVN
2
#
$
' (
∂ f3
∂ f3
1
Iyy Cm q
ρVN
2
#
$
' (
∂ f3
1
Iyy Cmα
ρVN
2
#
$
' (
• Coefficients in third row of F"
Sensitivity of Pitch Rate Dynamics to State Perturbations "
C m V ≡∂C m
∂V
C m q≡∂C m
∂q
C m
α ≡∂C m
∂α
q = Cm ρ V2
2
( ) Sc Iyy
Trang 5∂ f 4
∂V
∂γ
• Coefficients in fourth row of F"
Sensitivity of Angle of Attack
Dynamics to State Perturbations "
∂ q
∂α
α = θ − γ = q − γ
Dimensional Stability and Control Derivatives
Dimensional Stability-Derivative
Notation"
! Dimensional stability derivatives portray acceleration
sensitivities to state perturbations"
Drag mass (m) ⇒ D ∝ V
Lift mass ⇒ L ∝ V γ
Moment moment of inertia (Iyy) ⇒ M ∝ q
Dimensional Stability-Derivative
Notation"
∂ f1
m ( CT Vcos αN− CD V) ρ VN
2
2 S + C ( T Ncos αN− CD N) ρ VNS
&
'
* +
∂ f2
Lα
mVN ( CT NcosαN+ CLα) ρVN
2
%
&
)
*
∂ f3
Iyy Cmα
ρVN2
%
&
)
*
Thrust and drag effects are combined and represented by one symbol!
Thrust and lift effects are combined and represented by one symbol!
Trang 6Longitudinal Stability Matrix"
FLon = FPh FSP
Ph
FPh SP
FSP
!
"
#
#
$
%
&
& =
LV
VN
g
VNsinγN
Lq
− LV
g
VNsinγN 1− Lq
VN
* +
-.
VN
!
"
#
#
#
#
#
#
#
#
$
%
&
&
&
&
&
&
&
&
Effects of phugoid perturbations on phugoid motion"
Effects of phugoid perturbations on short-period motion"
Effects of short-period perturbations on phugoid motion"
Effects of short-period perturbations on short-period motion"
Primary Longitudinal Stability
Derivatives"
D V−1
m C T V − C D
V
( )ρV N
2
2 S + C T N − C D
N
( )ρV N S
#
$
' (
Assuming γN αN 0
L V
V N 1
mV N C L V
ρV N
2
2 S + C L N ρV N S
"
#
%
&
' −mV1
N
2 C L N
ρV N
2
2 S − mg
"
#
%
&
M q= 1
I yy C m q
ρV N2
2 Sc
"
#
%
&
Mα= 1
I yy C mα
ρV N
2
2 Sc
#
$
&
'
Lα
mV N (C T N + C Lα)ρV N
2
2 S
#
$
' (
Origins of Stability Effects
Velocity-Dependent Derivative
Definitions"
• Air compressibility effects are a principal source of velocity dependence"
CD
∂ ( V / a ) = a
C D V ≡∂C D
∂V =
1
a
#
$
&
'
(C D M
C L V ≡∂C L
∂V =
1
a
#
$
&
'
(C L M
C m V≡∂C m
∂V =
1
a
#
$
&
'
(C m M
C D M ≈ 0
C D M > 0 C D M < 0
a = Speed of Sound
M = Mach number = V a
Trang 7Pitch-Moment Coefficient
Sensitivity to Angle of Attack"
MB= Cmq Sc ≈ C ( m o+ Cm qq + Cmαα ) q Sc
Cm
αnet ( hcm − hcp net)
αnet
xcm − xcp net
c
$
%
(
αht
Pitch-Rate Derivative Definitions"
C m
q=∂C m
∂ q =
c
2V N
"
#
&
'C m ˆq
C m ˆq=∂C m
∂ ˆq =
∂C m
∂ qc 2V( N)=
2V N
c
"
#
%
&
'C mq
ˆq = qc 2VN
M q=∂M
∂ = C m q( ρV N2 2)Sc = C m ˆq c
2V N
#
$
% &
' ( ρV N 2 2
#
$
% &
'
(Sc = C m ˆq
ρV N Sc2 4
#
$
% &
' (
M B = C m q Sc ≈ C m
o+C m
≈ C m o+∂C m
∂q q + C mαα
$
%
(
)q Sc
• Pitch acceleration sensitivity to pitch rate "
Pitch-Rate Derivative Definitions"
in terms of a normalized pitch rate "
c
2VN
"
#
$ %
&
'Cm ˆq
C m ˆq=∂C m
∂ˆq =
∂C m
∂ qc 2V
N
2V N
c
"
#
$ %
&
'C m q
ˆq = qc 2VN
M B = C m q Sc ≈ C m
o+C m
≈ C m o+∂C m
∂q q + C mαα
$
%
(
)q Sc
• Then"
• But dynamic equations require ∂C m /∂q "
Angle of Attack Distribution
Due to Pitch Rate"
• Aircraft pitching at a constant rate, q rad/s, produces a normal velocity distribution along x"
Δw = −qΔx
Δ =Δw
V N
=−qΔx
V N
• Corresponding angle of attack distribution"
Δ ht=ql ht
V N
• Angle of attack perturbation at tail center of pressure"
€
l ht = horizontal tail distance from c.m.
Trang 8Horizontal Tail Lift
Due to Pitch Rate"
• Incremental tail lift due to pitch rate, referenced to tail area, S ht"
• Lift coefficient sensitivity to pitch rate referenced to wing area"
ΔLht = ΔC ( L ht)ht
1
2 ρ VN
2
Sht
C L qht ≡∂ ΔC( L ht)aircraft
∂C L ht
∂α
%
&
)
*
aircraft
l ht
V N
%
&
' ( )
*
ΔC L ht
( )aircraft = ΔC( L ht)ht
S ht
S
"
#
%
&
' = ∂C L ht
∂α
"
#
&
'
aircraft
Δ
* +
, ,
-
/ /=
∂C L ht
∂α
"
#
&
'
aircraft
ql ht
V N
"
#
$ %
&
'
• Incremental tail lift coefficient due to pitch rate, referenced to
wing area, S"
Moment Coefficient Sensitivity to Pitch Rate of
the Horizontal Tail"
∂ ΔM ht
∂ q = C m qht
1
2ρV N
2Sc = −C L qht
l ht
V N
%
&
( )
1
2ρV N
2Sc
= − ∂C L ht
∂α
%
&
( )aircraft
l ht
V N
%
&
( ) ,
0
1l ht
c
%
&
( ) 1
2ρV N
2
Sc
C m qht= −∂C L ht
∂α
l ht
V N
$
%
& ' ( ) l ht
c
$
%
' ( ) = −∂C L ht
∂α
l ht
c
$
%
' (
2
c
V N
$
%
& ' ( )
• Coefficient derivative with respect to normalized pitch rate is insensitive to velocity"
Cm ˆqht = ∂Cm ht
∂ ˆq =
∂Cm ht
∂ qc 2V ( N) = −2
∂CL ht
∂α
lht c
$
%
& ' ( )
2
Comparison of Fourth-
and Second-Order
Dynamic Models
• 0 - 100 sec "
4th-Order Initial-Condition Responses of Business Jet
at Two Time Scales"
• 0 - 6 sec "
• Plotted over different periods of time"
– 4 initial conditions "
Trang 9Second-Order Models of
Longitudinal Motion"
• 2 nd -Order Approximate Phugoid Equation"
ΔxPh= Δ V
Δ γ
#
$
%
%
&
'
(
(≈
−D V −g cosγ N
L V
V N
g
V NsinγN
#
$
%
%
%
&
'
( ( (
ΔV
Δγ
#
$
%
%
&
'
( (+
T δT
L δT
V N
#
$
%
%
%
&
'
( ( ( ΔδT +
−D V
L V
V N
#
$
%
%
%
&
'
( ( (
ΔV wind
ΔxSP= Δ q
Δ α
#
$
%
%
&
'
(
(≈
M q Mα
1 −L q
V N
+
,- /0 −
Lα
V N
#
$
%
%
%
&
'
( ( (
Δq
Δ
#
$
%
%
&
'
( (+
M δ E
−L δ E
V N
#
$
%
%
%
&
'
( ( (
Δδ E +
Mα
−Lα
V N
#
$
%
%
%
&
'
( ( (
Δ wind
• 2 nd -Order Approximate Short-Period Equation"
!
"
#
#
$
%
&
&
• Full and approximate linear models"
Comparison of Bizjet Fourth- and Second-Order Model Responses"
• Fourth Order"
• Second Order"
because natural frequencies are widely separated"
Comparison of Bizjet Fourth- and
Second-Order Models and Eigenvalues"
Fourth-Order Model
F = G = Eigenvalue Damping Freq (rad/s)
-0.0185 -9.8067 0 0 0 4.6645 -8.43e-03 + 1.24e-01j 6.78E-02 1.24E-01
0 0 -1.2794 -7.9856 -9.069 0 -1.28e+00 + 2.83e+00j 4.11E-01 3.10E+00
-0.0019 0 1 -1.2709 0 0 -1.28e+00 - 2.83e+00j 4.11E-01 3.10E+00
Phugoid Approximation
F = G = Eigenvalue Damping Freq (rad/s)
Short-Period Approximation
F = G = Eigenvalue Damping Freq (rad/s)
Approximate Phugoid Roots "
• Approximate Phugoid Equation (!N = 0)"
ΔxPh= Δ V
Δ γ
#
$
%
%
&
'
( (≈
−D V −g
L V
V N 0
#
$
%
%
%
&
'
( ( (
ΔV
Δγ
#
$
%
%
&
'
( (+
T δT
L δT
V N
#
$
%
%
%
&
'
( ( ( ΔδT
• Characteristic polynomial"
sI − FPh = det sI − F ( Ph) ≡ Δ(s) = s2+ DVs + gLV/ VN
= s2+ 2ζωns + ωn
2
• Natural frequency and damping ratio"
2 L / D ( )N
compressibility effects"
Trang 10Effect of Airspeed and L/D on Approximate
Phugoid Natural Frequency, Period, and
Damping Ratio "
ζ ≈ 1
2 ( L / D )N
Velocity
Natural
Damping Ratio
Neglecting compressibility effects"
Period, T = 2 π / ωn
≈ 0.45VN sec
Approximate Phugoid Response to
a 10% Thrust Increase "
• What is the steady-state response?"
Approximate Short-Period Roots "
• Approximate Short-Period Equation (L q = 0 )"
• Characteristic polynomial"
ΔxSP= Δ q
Δ α
#
$
%
%
&
'
( (
≈
M q Mα
V N
#
$
%
%
%
&
'
( ( (
Δq
Δα
#
$
%
%
&
'
( ( +
M δ E
−L δ E
V N
#
$
%
%
%
&
'
( ( (
Δδ E
Δ(s) = s2+ Lα
V N − M q
$
%
& '
(
)s − Mα + M q
Lα
V N
$
%
& '
( )
= s2+ 2ζωn s +ω n2
ωn = − Mα+ M q
Lα
V N
$
%
& '
( ); ζ =
Lα
V N − M q
$
%
' (
2 − Mα+ M q
Lα
V N
$
%
& '
( )
Generally,
Lα> 0
Mα< 0
M q< 0
Approximate Short-Period Response
to a 0.1-Rad Pitch Control Step Input "
Pitch Rate, rad/s! Angle of Attack, rad!
Trang 11Normal Load Factor Response to a
0.1-Rad Pitch Control Step Input "
Normal Load Factor, g s at c.m.!
Aft Pitch Control (Elevator)!
Normal Load Factor, g s at c.m.!
Forward Pitch Control (Canard)!
n z=V N
g (Δ α − Δq)=V N
g
Lα
V N
Δα +L δ E
V N
Δδ E
%
&'
( )*
• Normal load factor at the center of mass"
• Pilot focuses on normal load factor during rapid maneuvering"
Grumman X-29!
Next Time:
Lateral-Directional Dynamics
Reading
Flight Dynamics, 96-101, 574-582, 587-591 Virtual Textbook , Part 12
Supplemental Material
Flight Path Angle Dynamics"
• Second row of linearized equation"
γ = f2 = 1
mV[T sinα + L − mg cosγ]= 1
mV C TsinαρV
2
2 S + C L
ρV2
2 S − mg cosγ
%
&
( )
• Nonlinear equation"
Δ γ (t) = ∂ f2
∂ V ΔV (t) +
∂ f2
∂γ Δ γ (t) +
∂ f2
∂ f2
∂α Δ α (t)
%
&
)
*
+ ∂ f2
∂δ E Δ δ E(t) +
∂ f2
∂δ T Δ δ T (t) +
∂ f2
∂δ F Δ δ F(t)
%
&'
( )*
+ ∂ f2
∂ VwindΔVwind+
∂ f2
∂αwind
Δ αwind
%
&
)
*