aero dy namik

7/29/2019 Aero Dy Namik

1/20

Aerodynamik

Astrid Herbst (Bombardier), Tomas Muld & Gunilla Efraimsson ( KTH)


2/20

2

Future demands Higher speeds and high

capacity

Wide body trains with max speed > 250 km/h

Aerodynamic challenges specific for wide body trains:

head pressure pulse & slipstream


3/20

3

Slipstream

Safety issue for passengers on platform and trackside workers

Measures on operation & train design


4/20

More value for environment and performance:

Aerodynamic optimisation

Reduction of drag saves energy and traction power

Drag and Cross-Wind Optimisation


5/20


Aerodynamic optimisation

Parameterized model

defines the variables and

boundary conditions

Computer optimisation by

using the parameterizedmodel

Goal function is reduce

drag while keeping the

cross wind safety

chamfering

tangent between

nose and car body

upper part

of the nose tip

lower part

of the nose tip

starting section

of the nose

height and length

of the nose

size of the bogie fairing

spoiler geometrylateral tangent

at the nose tip

upper curvature

of the carbodyupper curvature

of the nose tip


6/20

Pareto front

Design Space

Lo

wdrag

Increasing stability

Pareto front

Design Space

Lo

wdrag

Lo

wdrag

Increasing stability


Aerodynamic optimization

Thousands of virtual wind tunnel

tests in the computer used to find

the very best shape

Main result shows 20 30 %

lower drag and 10 15 % lowerenergy consumption

Installed power can be reduced

with lower cost

Lower energy cost for operators


7/20

7

Slipstream Test Setup

Measurements with fast 3D-ultrasonic anemometers (USA)

Train speed and positions measured with light switch

Light switch mounted to detect single rail passages

Vs

ters

Enkp

ing

60 m CarCatenary Pole 1

Road

20 m 20 m 6 m 5 m

LS 1 LS 2

30.58 m

Catenary Pole 2

T-junction

9m

13m

Reference

Pressure

9.42 m

Up Track

Down Track

Vsters Enkping

USA 1 USA 2 HPP USA 3

Legend

USA : Ultra Sonic AnemometerHPP : Head Pressure Pulse Rake

LS : Light Switch

60 m

CarCatenary Pole 1

Road

20 m 20 m 6 m 5 m

LS 1 LS 2

30.58 m

Catenary Pole 2

T-junction

9m

13m

Reference

Pressure

9.42 m

Up Track

Down Track

Vsters Enkping

USA 1 USA 2 HPP USA 3

Legend

USA : Ultra Sonic AnemometerHPP : Head Pressure Pulse Rake

LS : Light Switch


8/20

8

Measurement results

-5

-2,5

0

2,5

5

7,5

10

-50 0 50 100 150 200

x [m]

u Wind[m/s]

-0,072

-0,036

0

0,036

0,072

0,108

0,144

cu,Wind[]

Ensemble average overall

Ensemble average Skirt leading

Ensemble average Skirt trailing

Maximum

Axles

Intercar

Gap

Head

Passage

Impact of

Bogie Skirts

Tail Passage


9/20

Numericalsimulations of slipstream

performed at KTH


10/20

Slipstream

Slipstream= Inducedvelocity by the train

Regions

1) Head pressure pulse

2) Boundary Layer(Freight Trains)

3) Near wake(High-speed Train)

4) Far Wake


11/20

Train models

Aerodynamic Train Model (ATM) Regina (CRH1)


12/20

Decomposition models

=

+= + +

Mode 1 Mode 2,3 Mode 4,5 Mode 6+

Full

flow

field


13/20

Connection between modes

Phase portrait

Spiraling circles when the modes are connected

Random patterns when not


14/20

Example of flow structure

Isosurface of

V-velocity

Mode 1+4+5


15/20

Decomposition models

Challenges

- Long computation times

- Accurate results

Understanding of dominant energetic structures

Two methods

- Proper Orthogonal Decomposition (POD)

- Dynamic Mode Decomposition (DMD)


16/20

Comparing with Experiments

CFD

by KTH

Water

Towing Tank

from

Bombardier

performed atDLR


17/20

Grid study

Small cells

Medium cells

Large cells


18/20

TSI-measurements3 m

1.2 m

Velocity for an observer

standing on platform

0.07s (1s) time

averaged velocity


19/20

Achievements

Showed that POD and DMD can be used with Detached EddySimulation flow fields.

Identified dominant flow structures for two different trains.

Used advanced models on applied geometries.

Cooperation and exchange of knowledge with Bombardier.

Fundament for efficient prediction to connect train geometryand slipstream.


20/20

Publications Muld, T.W, Analysis of Flow Structures in Wake Flows for Train Aerodynamics, Licentiate Thesis in Mechanics, KTH

Stockholm, ISBN 978-91-7415-651-5, 2010

Muld T.W, Efraimsson G., Henningson D. S, Flow structures around a high-speed train extracted using ProperOrthogonal Decomposition and Dynamic Mode Decomposition, Computer & Fluids, DOI:10.1016/j.compfluid.2011.12.012 , 2012

Muld T.W, Efraimsson G., Henningson D. S, Mode decomposition on surface mounted cube, Flow, Turbulence andCombustion, DOI: 10.1007/s10494-011-9355-y, 2011

Muld T.W, Efraimsson G., Henningson D. S, Mode Decomposition of Flow Structures in the wake of Two High-SpeedTrains, The First International Conference on Railway Technology: Research, Development and Maintaince, April 16-18,Gran Canaria, Spain, 2012

Muld T. W., Efraimsson G., Henningson D. S., Herbst A. H. and Orellano A., Analysis of Flow Structures in the Wakeof a High-Speed Train, Aerodynamics of Heavy Vehicles III: Trucks, Buses and Trains, September 12-17, 2010,Potsdam, Germany

Muld T. W., Efraimsson G., Henningson D. S., Herbst A. H., Orellano A., Detached Eddy Simulation and Validation onthe Aerodynamic Train Model, Euromech Colloquim 509, Vehicle Aerodynamics, Berlin Germany, March 24-25 2009

Muld T. W., Efraimsson G., Henningson D. S., Proper Orthogonal Decomposition of Flow Structures around a Surface-mounted Cube Computed with Detached-Eddy Simulation, SAE paper 09B-0170, 200915

aero dy namik

Documents