Data Assimilation Method for EFD/CFD IntegrationBoeing/ppt/201411/BHE_Report_2014_ambo.… · Data Assimilation Method for EFD/CFD Integration 2014/11/20 1 Takumi Ambo, Chen Fang

Data Assimilation Method for

EFD/CFD Integration

2014/11/20 1

Takumi Ambo, Chen Fang

Department of Aerospace Engineering / Institute of Fluid Science

Tohoku University

Boeing Higher Education Program Presentation


2014/11/20 2

Background -Aerodynamics of low Re number condition-

Appearance of Small airplanes

Unmanned air vehicle (UAV)

Micro air vehicle (MAV)

Three-dimensionality Generation of wingtip vortex

Flight in Low Re conditions (Re = 104~105)

Aerodynamic characteristics

• Different from High Re number condition

• Nonlinearity

Flow field around wing

• Separation or transition of boundary layer

• Formation of separation bubble

It is necessary to investigate aerodynamic characteristics

on the three-dimensional wing in low Re condition.

Three-dimensional flow field

Schematic diagram of separation bubble

(Muller)

Lift coefficient of NACA0012-34 airfoil

(Anyoji et al，2010)

-0.4

-0.2

0

0.2

0.4

0.6

0.8

-5 0 5 10 15

Re=1.1x104, M=0.21

Cl

[deg]

Unmanned Air Vehicles

(http://en.wikipedia.org/wiki/Unmanned_aerial_vehicle)


Background -EFD & CFD-

2014/11/20 3

Experimental Fluid Dynamics

(EFD)

Computational Fluid Dynamics

(CFD)

Advantage Advantage

Credible data

(Real flow)

Low cost

Big information content

Disadvantage

Disadvantage High cost

Long experimental time

Limited information

Difference from the real flight

(Wall & Support interference)

Computational accuracy relies

on the scheme and mesh

Complement


2014/11/20 4

CFD/EFD

(Data Assimilation)

CFD

(Forecast)

EFD

(Observation)

Data Information

Quality Control

Validation

Forecast error

First Estimated Value

Model Improvement

Background -Data Assimilation-

• Data assimilation is an analysis technique in which the observed

information is accumulated into the model state by taking advantage of

consistency constraints with laws of time evolution and physical properties.


Past Studies

Data assimilation

• Computational Study of Flow around a Delta Wing for Measurement-

Integrated Simulation with Dynamic Wind Tunnel Testing (Shuta Ito, 2013)

2014/11/20 5 Boeing Higher Education Program Presentation

Objective

2014/11/20 6

Make the numerical simulation become high-

precision using Data assimilation method to

understand the aerodynamic phenomenon

Contents

EFD

Calculating Lift coefficient and surface streamline

Force measurement

Flow visualization

CFD


EXPERIMENTAL FLUID

DYNAMICS (EFD)

2014/11/20 Boeing Higher Education Program Presentation 7

Experimental Equipment

2014/11/20 8

Wind-tunnel

Test model

Low Turbulence Wind Tunnel (Tohoku Univ.)

- Type : closed circuit

- Test section : open type and octagon

- Test section size : cross distance of 0.8 m

- Turbulence intensity : 0.02 %

Wing

- Airfoil : NACA0012

- Chord length : 75 mm

- Span length : 400 mm

- Material :Wood(MDF)

14

0 m

m

400 mm

75

mm

φ 30

Body

- Diameter : 30 mm

- Material : Aluminum


400 mm

φ30

75 m

m

NACA0012 Airfoil

2014/11/20 9

Measurement method -Force measurement-

Sting six-component balance

(NISSHO-ELECTRIC-WORKS, LMC-61112)

Rated load

・Fx－120 N

・Fy－250 N

・Fz－250 N

Sting six-component balance

LFMM zyc 4/

Calculating Pitching-moment

・Mx－4 N・m

・My－10 N・m

・Mz－10 N・m

1/4 chord length Moment center


Global Luminescent Oil-Film Skin-Friction

Meter (GLOF method)

• Oil doped with luminescent molecules is put on

the test model.

• Skin friction can be measured to capture time-

changing of emission intensity.

Systems

• Fast camera ： Lends and optical filter are set

forward of camera

• Light source：UV-LED units

Luminescent oil film

• Oil：Silicone oil

• Fluorescent dye：Perylene

2014/11/20 10

Measurement method -Visualization-

Fast camera

UV-LED

Model

Skin friction measurement

Global Luminescent Oil Film [GLOF]

Measurement system

・ Fast camera ： Lends and optical filter are set forward of camera

・ Light source：UV-LED units

Luminescent oil film

・ Oil：Silicone oil

・ Fluorescent dye：Perylene

・ Oil doped with luminescent molecules is put on the test model.

・ Skin friction can be measured to capture time-changing of emission intensity.

Luminescent oil film on the model -1 -0.5 0 0.5 1

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

y/cx/c

Line of frictional stress


COMPUTATIONAL FLUID

DYNAMICS (CFD)


Computation


𝐷𝜌𝑘

𝐷𝑡= min 𝑃,

𝑅𝑒∞𝑀∞20𝛽∗𝜌𝜔𝑘 −

𝑅𝑒∞𝑀∞𝛽∗𝜌𝜔𝑘 +

𝑀∞𝑅𝑒∞

𝜕

𝜕𝑥𝑗(𝜇 + 𝜎𝑘𝜇𝑡)

𝜕𝑘

𝜕𝑥𝑗 𝛾 ∗

Realization of the laminar flow - turbulence transition calculation

： Turbulence energy equation of Menter SST model

Intermittency factor distribution

(Laminar flow cell γ = 0 , Turbulent flow cell γ = 1)

FaSTAR : Developed by JAXA

Aerodynamic coefficients

Surface streamline

Number of ：200万点

Flow

Laminar

Turbulent

Computation


Flow

＊Initial condition : It is decided based on separation position of experimental results.

Assimilation method：Ensemble Kalman Filter (EnKF)

Initial condition Parameter Convergent calculation Filtering Calculation end

The number of the

ensemble ：10

Experimental

value

Observation update

Laminar

Turbulent

Control point

Computation


𝑎𝑦

𝑎𝑥

𝑏𝑦

𝑏𝑥

Difference of observation and model :

𝐻 𝑥 − 𝑦 =𝑎𝑦

𝑎𝑥−𝑏𝑦

𝑏𝑥

Objective function of Kalman filter

(minimization of 𝐽):

𝐽 = 1

2(𝑎𝑦

𝑎𝑥−𝑏𝑦

𝑏𝑥)2

𝑆𝑢𝑟𝑓𝑎𝑐𝑒𝑚𝑒𝑠ℎ 𝑝𝑜𝑖𝑛𝑡

Same position

Computational grid Line of frictional stress

(Experiment)

Test conditions


Wind-tunnel Low turbulence wind-tunnel

Free-stream velocity [m/s] 30

Angle of Attack, α [deg] -10 ~ 30 [deg]

Measurement Aerodynamic forces, Frictional stress line

Experimental conditions

Free-stream velocity [m/s] 30

Angle of Attack, α [deg] 0 ~ 15

Re 1.5 × 105

Turbulence model Menter SST

Analysis method Steady

Number of grid 2× 106

Calculation conditions

RESULTS


Results -Lift coefficient-

Lift force

2014/11/20 17

-0.4

-0.2

0

0.2

0.4

0.6

0.8

-5 0 5 10 15 20

Re=2.5x104

Re=5.0x104

Re=1.5x105

CL

Angle of Attack [deg]

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

-4 -2 0 2 4

Re=2.5x104

Re=5.0x104

Re=1.5x105

CL

Angle of Attack [deg]

低Re数領域においては，揚力特性はRe数に大きく依存する．

Re数の減少に伴って揚力傾斜の迎角に対する非線形性が強くなる．

Re=5.0x104以下では，低迎角域で特異な非線形領域が現れる．

Re数の減少に伴い低迎角域における非線形領域は拡大する．

The characteristics of lift force is different from high Re number condition.

Stall angle and maximum lift coefficient are lower.


-0.4

-0.2

0

0.2

0.4

0.6

0.8

-5 0 5 10 15 20

CL

Angle of Attack [deg]Theory of wing sections (1959)

Re = 1.5 x105

Results - Frictional stress line -

2014/11/20 18

2y/b

x/C

0 0.5

10.

5

1

0

2y/b

x/C

0 0.5

10.

5

1

2y/b

x/C

0 0.5

10.

5

2y/b

x/C

0 0.5

10.

5

1

0

2y/b

x/C

0 0.5

10.

5

1

0

2y/b

x/C

0 0.5

10.

5

1

2y/b

x/C

0 0.5

10.

5

2y/b

x/C

0 0.5

10.

5

1

0

2y/b

x/C

0 0.5

10.

5

1

0

2y/b

x/C

0 0.5

10.

5

1

2y/b

x/C

0 0.5

10.

5

2y/b

x/C

0 0.5

10.

5

1

0

α = -1.5 deg α = -0.5 deg α = 0.5 deg α = 4 deg

α = 8 deg α= 10 deg α = 12 deg

Laminar separation is occurred near trailing edge.

Separation bubble appears after α=-0.5 deg.

Separation bubble shift forward along with

increase of AoA.

Leading edge separation bubble is formed in α=8

deg.

Flow separate from trailing edge in α=10 deg.

Most of flow field is dominated by reverse flow

is α=12 deg.

α = 5 deg

Separation line Reattachment line

Separation

bubble


Root

Wing tip

Flow

Results -Calculated Lift coefficient-


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 2 4 6 8 10 12 14 16 18

Lif

t co

effi

cien

t, C

l

Angle of Attack, α [deg]

Experiment

SST

Laminar

-0.1

-0.2

-0.3

cp5

Under α= 8 deg ：Experimental value ≈ Calculated value.

After α = 8 deg There’re difference between EFD and CFD results.

High AoA Separation bubble in leading edge Convergence is difficult.

Shifting boundary of laminar and turbulent Cl changes.

Results - Calculated Surface streamline, α = 8 deg -


Experimental frictional stress line

Without transition

0

-0.5

-1

Flow

Transition line locates x/c = 0.1 Transition line locates x/c = 0.2

Transition line locates x/c = 0.3 Transition line defined with 5 control points

To control boundary layer transition points using data assimilation method is

useful as a means for approximating CFD results with EFD results

Summary

• In EFD, force measurement and flow visualization were conducted.

• In CFD, location of transition line influences lift coefficient and

surface streamline.


Future Works

• Conducting calculation of controlling boundary layer transition points

using data assimilation method.

• Apply this method to other unsteady flow conditions like dynamic wind-

tunnel testing.

Make the numerical simulation become high-precision using Data

assimilation method to understand the aerodynamic phenomenon Experiment and Calculation using same model

were conducted and each results were compared.

2014/11/20 22

Thank you for your attention


Documents

Data Assimilation Method for EFD/CFD IntegrationBoeing/ppt/201411/BHE_Report_2014_ambo.… · Data Assimilation Method for EFD/CFD Integration 2014/11/20 1 Takumi Ambo, Chen Fang