Validation of Wake Vortex Encounter Simulation Models Using Flight Test Data

Transcription

1 Validation of Wake Vortex Encounter Simulation Models Using Flight Test Data Assessment of Wake Vortex Safety Dietrich Fischenberg DLR Braunschweig Workshop WakeNet2-Europe, Working Group 5, Hamburg, -11 May 2004

2 2 Scope of Presentation flight tests, measurements, and data flight test data analysis performed within the S-WAKE project: determination of vortex model parameters to characterize vortex flow field validation of flight mechanic/aerodynamic interaction models summary

3 Statistics of Full Scale Flight Encounter 3 Date Encounter A/C Altitude Encounter flown Flap setting of wake generating A/C DLC-flap setting Dornier Do128 FL const Dornier Do128 FL , 35 const Dornier Do128 FL , 1 split, oscillating Cessna Citation FL const Dornier Do128 FL , 35 const. total 116

4 Wake Generating Aircraft ATTAS 4 ATTAS with extended flaps smoke generator in action smoke trace, constant DLC flaps smoke trace, oscillating DLC flaps

5 Encounter Aircraft 5 Dornier Do 128 (TU-BS) 4 flow probes (5 hole probes) 4,2 m 2,3 m Cessna Citation II (NLR) 1 flow probe (vanes)

6 Encounter Maneuver 6 wake generation 0.5 nm 1.5 nm 3.0 nm smoke trace

7 Flight Test Encounter Scenario 7 wake generation 0.5 nm 1.5 nm 3.0 nm smoke trace

8 Do128 Typical Encounter Flow Sensor Measurements 8

9 The 2 Steps of Encounter Flight Test Data Evaluation 9 step 2 measured flight flight test test data data pilot s control inputs basic basic A/C A/C aero aero model model aerodynamic interaction model model forces, moments + + forces, moments accelerations, s, attitude, altitude, velocity 6-DOF 6-DOF A/C A/C simulation simulation outputs - + model accuracy step 1 wake vortex characteristics step 2: encounter model validation accelerations, s, attitude, altitude, velocity flow measurements flight flight path path reconstruction reconstruction (FPR) (FPR) Vortex model step 1: flow field characterization

10 Validation Basic Do128 Model - no Wake Vortex Influence measured model output lateral accel. vertical accel. roll pitch 6 M/S M/S time, s yaw 40 bank -0 pitch yaw time, s

11 Determination of Vortex Model Parameters: Rosenhead - B & H 11 noseboom sensor horizontal velocity vertical velocity measured model output right wing sensor left wing sensor distance: 0.8 nm Identified: Γ = m 2 /s vertical tail sensor r C = 0.90 m (=4.2% wing span) time, s time, s

12 12 Determination of Vortex Model Parameters: Winckelmans horizontal velocity noseboom hit of vortex core sensor vertical velocity measured model output right wing sensor left wing sensor distance: 0.6 nm Identified: Γ = 152 m 2 /s r C = 0.22 m (=1% wing span) time, s time, s

13 Do128: Comparison of Different Vortex Models 13

14 14 Aerodynamic Interaction Models Strip Method (SM) ONERA Lifting Surface Method (LSM) TU Berlin

15 Validation of Strip Method (SM): Simulation of ATTAS/Do128 Wake Vortex Encounters 15 lateral accel. vertical accel. roll pitch 3 M/S2-3 0 M/S Encounter 1: right left Encounter 2: left right time, s yaw bank pitch yaw Encounter 1: right left Encounter 2: left right time, s measured simulation model output

16 Validation of Lifting Surface Method (LSM): Simulation of ATTAS/Do128 Wake Vortex Encounter 16 lateral accel. vertical accel. roll pitch 3 M/S2-3 0 M/S Encounter 1: right left Encounter 2: left right time, s yaw bank pitch yaw Encounter 1: right left Encounter 2: left right time, s measured simulation model output

17 Summary 17 S-WAKE flight test measurements (116 encounter) are a valuable high quality data base Flight test data were successfully evaluated with parameter identification and flight path reconstruction techniques to determine parameters of wake vortex models (Rosenhead & Burnham-Hallock, Lamb-Oseen, Winckelmans) Flight test data were successfully evaluated to validate aerodynamic interaction models (AIMs) for near parallel encounter cases (strip method, lifting surface method) In general, both AIMs are suitable to simulate wake vortex encounters (especially roll and vertical axes). Overall, both methods show equally good results.