Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 1 Aerodynamic Modeling and System Identification from Flight Data – Recent Applications at DLR TÜBITAK-SAGE / METU, Ankara, Turkey, 29 May 2008 by Dr. Ravindra Jategaonkar Senior Scientist Institute of Flight Systems DLR - German Aerospace Center Lilienthalplatz 7, 38108 Braunschweig, Germany Email: [email protected]Phone: 0049 531 295-2684 ?
47
Embed
Aerodynamic Modeling and System Identification from Flight ... · PDF fileAerodynamic Modeling and System Identification ... selected flight vehicle model ... - Flight control law
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 1
Aerodynamic Modeling and System Identification from Flight Data – Recent Applications at DLR
Examples of Aerodynamic Modeling- C-160 Aerodynamic Data Base:
Data consistency checkingNonlinear control surface effectivenessSeparation of pitch damping derivativesStall hysteresisModeling of landing gear effects
- Database validation- Wake Vortex Aircraft Encounter Model- EC-135 Helicopter
6 DOF and extended models and Rotor wake modeling- Phoenix: Reusable orbiter glider, - UAV: Automatic Envelope Expansion through Adaptive Flight Control
Concluding remarks
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 3
Dynamic System u z
Aircraft masscharacteristics
Aerodynamics(unknownparameters)
Sensorlocations
Sensor model(calibration factors,bias errors)
Inputs States
Process noise(turbulence)
Measurementnoise
Outputs
State Equations Measurement Eq.)),(),(()( Θ= tutxgty)),(),(()( ϕtutxftx =&
AIM: To determine unknown model parameters Θ such that the model response y matches well with the measured system response z.
What is System Identification? (1)
yx
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 4
Classification
Simulation: given u and f, find x
Control: given x and f, find u
Identification: given u and x, find f
Inputs Outputs
u xState Equations
x = f (x, u, θ).
Simulation
Parameter estimation System
identification
Given the answer, what are the questions, i.e., look at the results and try to figure out what situation caused those results.
SysID: an Inverse Problem
(1) System IdentificationConcerned with the mathematicalstructure of a flight vehicle model
(2) Parameter EstimationQuantifying of parameters for aselected flight vehicle model
Is the commonly used terminology PID appropriate?
?
What is System Identification? (2)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 5
Need and quest to better understand the system- Cause-effect relationship purported to underlie the physical phenomenon
Mathematical models required for:- Investigation of system performance and characteristics
- Aerodynamic databases valid over operational envelope for flight simulators
- High-fidelity / high-bandwidth models for in-flight simulators
- Flight control law design
- Analysis of handling qualities compliance
Aerodynamic databases from flight data- Analytical estimates: validity and inadequate theory !
- Wind-tunnel predictions: model scaling, Reynold's number,
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 21
Vertical-acceleration
Pitchrate
Pitchattitude
Pitchacceleration
CG-location
-8
-135
-5
8
-28
-16
50
20
-8
-135
-5
6
28
-16
50
20
m/s2m/s2
deg/s deg/s
degdeg
2deg/s 2deg/s
% %
0 5 10 15Time (sec) 0 5 10 15
Neglecing Variations in Aircraft MassCharacteristics
Accounting for Variations in theAircraft Mass Characteristics
MeasurementSimulation
Time (sec)
C-160: Load Drop (4.6 t)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 22
How do you know that you got theright answer?
1.2.3.4.
5.
Standard derivationsCorrelation among the estimatesGoodness of fitPlausibility of estimates(WT data base)Model predictive capability
"ACID TEST"Simulation and comparison withflight data not used in identification
Time, sec
-900.1 10
-1090
0
-12
-70 2 4 6
Frequency, rad/sec
Magn.
Phase
deg
deg/secq
10
9deg
δe
10dB
0
modelmeasured
qq
1001.0
modelmeasuredq
q
C-160 Pitch Response
model
modelq
Data Base Validation (1)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 23
δf = 30°
δf = 0°
-10 150angle of attack
deg
Cnβ
0.1
0.2
0.31/rad WT/analytical
prediction
flight estimates
Weathercock stability
50 10 15 20time s
10
-10
0
8
-88
-8
0
0
deg/s
deg
deg
yawrate
angle ofsideslip
rudder
Flight measured
WT database prediction
SysID database prediction
Dutch roll dynamics
WT-Predictions: 4.18 s 0.207
Flight estimated Database: 5.04 s 0.202
Flight recorded responses: 5.12 s 0.198
Tolerances:
Frequency: +- 0.5 s or 10%
Damping: +- 0.02
Data Base Validation (2)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 24
Do-328: Stand-alone versus Integrated Models
ReversibleFlight Control
Dynamics
Aircraft MotionVariables
Pilot InputForces
Control SurfaceDeflect.
Aircraft MotionVariablesRigid Body
Dynamics
Flight controls stand-alone
ReversibleFlight Control
Dynamics
Rigid BodyDynamics
Integrated model
Control SurfaceDeflect.
Aircraft MotionVariables
Pilot InputForces
Control SurfaceDeflect.
Rigid-body stand-alone
measured data simulated data
Data Base Validation (3)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 25
DO-328:Normal LandingData Base Validation (4)
Flight measured (DO 328)Model identifiedFAA AC 120-40C tolerances
δe
Expanded View (Touch down)
θ
φ
14 16
-5
52
80
60
0
-10
ft
deg
deg
deg
h
Time (sec)
1.5 deg
2 deg
10 ft
18
130
0 10 20-5
10-10
20-10
5-5
100
25090
kt
ft
deg
deg
deg
deg
V
h
θ
φ
δe
δr
1.5 deg
10 ft
3 kt
2 deg
Complete Landing Phase
Time (sec)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 26
Engine failure during the critical phase of takeoff
Response to rudder and aileron important
Complete sequence as a single timesegment (stand-still, acceleration,Rotation, and climb to 200 ft)
No closed-loop controller
Tolerances: 3 kt on airspeed20 ft on altitude1.5 deg on pitch attitude2.0 deg on bank
Validation Example 3: Critical Engine Failure (DO 328)
Data Base Validation (5)kt
ft
deg
deg
deg
deg
150
-500
300
-1000
15
-50
-5
0
10
30
-15
0
10
-20
0
0
-2000
4000daN
0 10 20 30 stime
left engine shut off
V
h
θ
φ
δe
δr
FL , FRFlight measuredModel identified
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 27
Full Scale Flight tests: Data Gathering
Wake Vortex Aircraft Encounter Model (1)
smoke trace
0.5 nmwake
generation1 nm
1.5 nmWake vortex encounters:
- Aircraft reaction dominantly affected - Critical situation during safety-critical flight phases (landing, vicinity of airport)
Full scale flight tests with ATTAS followed by Do-128 or Cessna Citation
Separation class medium
100 encounters under steady atmospheric conditions.
Reaction of follower Aircraft:- Up to 80° bank angle; typical 30-40°; Bump; Uncomfortable; usually does notlead to loss of control (banking motion is averaged out)
- More important: lateral acceleration; may lead to injury to crew or Passenger
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 28
Wake Vortex Aircraft Encounter Model (2)
Schematic of Two Step Procedure for Vortex Model Identification
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 29
Wake Vortex Aircraft Encounter Model (3)
Analytical Model
rCy
bv
rightvortex
leftvortex
Vt,maxVt
rC
⎟⎠⎞⎜
⎝⎛ −
Γ=
−
− 22 /2544.112
)(
:
crrt e
rrV
OseenLamb
π222
)(
:
rcr
rrtV
HallockBurnham
+
Γ=
−
π
The model consists of two idealized, superimposed counter-rotating single Vortices. Model parameters:- vortex circulation Γ, - core radius rC, - lateral vortex separation bV, - vortex location in space.
The tangential velocity of one vortex as a function of the distance from thecore, Vt(r), is described in terms of the circulation Γ and the core radius rC:
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 30
Wake Vortex Aircraft Encounter Model (4)
Wake velocity components during lateral encounter
-15
10
-8
-8
6
6
m/s
m/s
m/s
0 2 4 6time S0 2 4 6S
horizontal velocity vertical velocityN
oseb
oom
Rig
ht w
ing
Left
win
g
m/s
m/s
m/s
-20
15
-15
-1510
10
time
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 31
Identified core radius rCand circulation Γ of the Burham-Hallock model for do-128 encounters from three flights.
Conformance with Theory:
- Expected decay of circulation
- Increase of core radius
- Initial core radius ~ 0.75 m, (roughly 3.5% of the wakeGenerating wing span which Is somewhat smaller thanCommonly stated value of 5%)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 32
0 10 20 30 40time
s
rad/s
rad/s
rad/s
rad
rad
0.3
-0.50.3
-0.3
0
0
0.3
-0.3
0
0.2
-0.2
0
0.3
-0.3
0
75
35
%
p
q
r
α
β
δlat,δped
Model Predictive Capability
EC-135 Flying Helicopter Simulator (1)
Forward speed 60 kts:Two flight maneuvers (Lateral and pedal inputs)
6-DOF Rigid-Body model:- Angle of attack dependent lateral-directional derivatives- Nonlinear aerodynamics; Weathercock stability for +ve and -ve sideslip angles
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 33
Rotor Wake ModelingRoll and pitch in hover and at low speeds:
unsymmetrical vortex compression anddilatation act on the induced velocity fieldin the proximity of the main rotor.
Effective AoA at the blade sections changed.
Aerodynamic rotor loads directly affected.
Rotor gyroscopic behavior due to the bladeflapping dynamics forced by these loadsleads to strong cross coupling effects of thehelicopter due to the wake distortion.
Current research topic:Suitable flight dynamic models describingthis phenomenon to obtain improvedsimulation fidelity in off-axis response
Pure Hover
Pitching motion in Hover
EC-135 Flying Helicopter Simulator (2)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 34
M: Apparent mass matrix associated with theacceleration terms from momentum theory
L: gain matrix, λ (= [λ0, λs, λc]T) the inflow ratiodescribing the first harmonic terms
c (= [cT, c1, cm]T): rotor load coefficients wrt rotor thrust and aerodynamic pitch and roll moment,Ω: main rotor rotation speedKp and Kq: Wake distortion parameters for longitudinal and lateral distribution of the induced velocity.
Last term on RHS: Parametric term that feeds back the roll and pitch rates of the rotor tip path plane wrt to the surrounding air to the induced velocity distribution over the rotor disk
Estimate Kp and Kq
⎥⎥⎥⎥
⎦
⎤
⎢⎢⎢⎢
⎣
⎡
β−
β−−
Ω+=λ
−+λ
)cq(qK
)sp(pK0
1L1c1LM&
&&
Dynamic Wake Model: Parametric extension of Pitt and Peters:
EC-135 Flying Helicopter Simulator (3)
Theoretical estimates of Wake distortion parameters
From flight tests applying SysId methods: Kq = 1.6; Kp = 2.5 (μ = 0)μ= VH/ΩR; VH: forward speed m/s; Ω: main rotor rotation speed rad/s; R: rotor radius in m.
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 35
Lateral Inp
utLo
ngitu
dina
lInput
Roll Rate
Pitch Ra
te
deg/s
deg/s
deg
20
0
-30
30
0
20
-20
2
0
-3
time20 25 30 35
-20
deg4
0
-1
2
s
Flight case #1 Flight case #2with theoreticalWD-parameters
with flightidentified WD parameters
Simulation fidelity at HoverDynamic Wake Model: Parametric extension of Pitt and Peters:
EC-135 Flying Helicopter Simulator (4)
20
10
0
-10
-20
60 64 68stime
deg/s
PitchRate
Inflow Dynamics
20
10
-10
-20
60 64 68stime
deg/s
Inflow and WakeDistortion
0
Flight DataSimulation
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 36
with PWD
without PWD
time
Lateral Inp
utLo
ngitu
dina
lInput
Roll Ra
tePitch Ra
te
deg/s
deg/s
deg
deg
2
15
0
-10
20
0
-20
-40
-1
-3
-5
4
00 2 4 6 8 10s
Flight case #1 Flight case #2
EC-135 Flying Helicopter Simulator (5)
Dynamic Wake Model: Parametric extension of Pitt and Peters:
Forward speed 40 m/s
μ= VH/ΩR = 0.18
Theoretical estimate = 0
From flight tests applying SysId methods: Kq = 1.6; Kp = 1.1 Good match
But, estimates do not conform toThe wake distortion theory.
Anomaly: Parameters do notRepresent wake distortion which occurs at hover. They account forOther unmodeled effects (rigid / elastic blade formulation).
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 37
Wind-tunnel testing in August 2003
Pre-flight checks: April 2004
calibration of flow angles:
α-error nonlinear: quadratic or piecewise linear
Accuracy:AoA and AoS: < 0.5°Horizontal velocity: 0.5 m/s
Phoenix: Reusable orbital glider (1)
-10 -5 0 5 10 15 20 25-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
alpha
Alph
a-er
ror
afte
rlin
ear c
alib
ratio
n
40 m/s70 m/s100 m/s
deg
deg
offsetc
d
c
dqp
korrqK
pα++=α β
βα
α
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 38
Flight phases upon release:1) Acquisition2) Approach3) Flare4) Alignment5) Derotation6) Rollout
Launch at40m/s EAS
flare
118m/s
runway
acquisition dive
altitude
RLVapproachpath -23o
510m
2.65km6.6km
touch down71m/s
Towed to establish initial conditions
Phoenix freeflight profile
45m x 2100m
release Phoenixfrom helicopter
touchdown
lift-off
ground track
Launch at40m/s EAS
flare
118m/s
runway
acquisition dive
altitude
RLVapproachpath -23o
510m
2.65km6.6km
touch down71m/s
Towed to establish initial conditions
Phoenix freeflight profile
45m x 2100m
release Phoenixfrom helicopter
touchdown
lift-off
ground track
Phoenix: Reusable orbital glider (2)
Reference Mission:
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 39
Free flights:Maiden flight on 8-May-2004Repeat flight on 13-May-20043rd flight with Offset on 16-May-2004
Configuration:Delta Wing, relatively low wing span3 controls (flaperons and rudder)Body flap and speed brake1200 Kg7m long3,48 m span
Highly dynamic behaviorHigh bandwidth control loops
VideoFlight 1 and Flight 3
Phoenix: Reusable orbital glider (3)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 40
Aerodynamic Database: Verification and Update -- Principle
Phoenix: Reusable orbital glider (4)
MeasuredAX, AY, AZ,p, q, r,pdyn
p_dot, q_dot, r_dot
AX_cg, AY_cg, AZ_cg
X, Y, Z, L_cg, M_cg, N_cg
X, Y, Z, L_ac, M_ac, N_ac
Flight derivedCX, CY, CZCLX, CMY, CNZ
Windtunnel Database(aerodataV31)
Measureddero, delo,dari, deli,dr, dbf, dsb,p, q, r,al, be
WT-PredictionsCX, CY, CZCLX, CMY, CNZ
Δs
SysID
Corrections
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 41
Flight derived and WT predicted vertical force coefficient
Rough order of discrepancies:CZ: 9-10%
Cm: < 3%
CD: ~10% Nonlinear
Pre-flight ADB
Flight-derived
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 42
sbsbref
q CXVL
qCXCXCXCX δα δα +++=Δ 0
bfbfref
q CZVL
qCZCZCZCZ δα δα +++=Δ 0
sbsbee CMCMCMCMCMY δδα δδα +++=Δ 0
12 Parameters CZ(), CX() and Cm() are estimated to reduce the deviations between flight measurements and WT-predictions.
Aero model update (In-Air)
Phoenix: Reusable orbital glider (6)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 43
Delta CZ versus AoAwithout and with update
Important Inferences:- lift generated in flight is higher
- component due to pitch rate in lift and drag is not adequately accounted for.
- basic longitudinal force coefficient for clean configuration underestimated,
- impact of speedbrakes overestimated.
0 0.05 0.1 0.15 0.2 0.25-0.06
-0.05
-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
Angle of attack (rad)
Del
ta C
Z
ff-n1ff-n2ff-n3
0 0.05 0.1 0.15 0.2 0.25-0.06
-0.05
-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
Angle of attack (rad)
Del
ta C
Z
ff-n1ff-n2ff-n3
Flight derived and Updated databasePhoenix: Reusable orbital glider (7)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 44
Automatic Envelope Expansion through Adaptive Flight Control (1)
Déjà-vu: Modules of adaptive control system
Limitations
Reference Dynamics
Reference States
Reference Model
Coupling
Nonlinear Effects
Known Disturbances
Feed Forward
Stability
Error Response Dynamics
Error Compensation
Feedback
Nonlinear Compensator
InputHidden
Output layer
Σ
Expansion to flight regimes not accounted for during the controller design (Design based on linear Hover model)
Flight demonstration of adaptive flight control
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 45
Flight test results: Forward flight (>10m/s) with midiARTIS
Time in s50 100 150 200 250
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
| Vel
ocity
Con
trolE
rror |
in m
/s
without adaptive Elementmean errorwith adaptive Elementmean error
Significant reduction in error coomanded speed
Reduction in the meanerror by factor of 10
0
Automatic Envelope Expansion through Adaptive Flight Control (2)
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 46
The Future
Prime areas of applications:Aerodynamic database generationsModeling of nonlinear aerodynamicsUnstable aircraft
New measuring techniques for air dataFlush air data sensorsoptical sensors
Real-Time parameter estimation is re-emerging (after seventies)
Full flexible aircraft models (integration of flight mechanics and structuralmodels) -- distributed mass models
Modeling and identification of UAVs, mAVs
Integrating System Identification and Computational Fluid Dynamics methodologies
Dr. Ravindra Jategaonkar RTO Mission at Tübitak-SAGE, Ankara, Turkey, 26-29 May 2008 pp. 47
Concluding Remarks
Unified approach based on Quad-M basics and various aircraft parameter estimation methods
Various examples covering global aerodynamic database, nonlinear effects,stall hysteresis, landing gear effects, load drop
Modeling of wake vortex encounter
Modeling of rigid-body and extended models for EC-135 helicopter
Modeling of Reusable orbital glider
Different aspects and examples of validation of identified models
Summary:SysID methods provide a well proven and highly sophisticated toolfor aerodynamic modeling from flight data.
Experience, engineering judgement and skill to interpret the modeling discrepancies and formulate them mathematically mainly limits the scopeof applications.