School of Aeronautical Engineering, Queen’s University Belfast

Turbulent Wind Flow over a High Speed Train

R K CooperSchool of Aeronautical EngineeringQueen’s University Belfast

Acknowledgement: To Network Rail WCML for permission toUse experimental data obtained by BMT Fluid Mechanics Ltd.

School of Aeronautical Engineering, Queen’s University Belfast

Cross wind effects on trains

Low mass passenger vehicles may be blown over.•Wind tunnel tests used to give force and moment data•Probability of overturning may be evaluated

Wind tunnel data shows large variability•Effect of turbulence?•Effect of embankment?•Effect of train motion?

Moving model experiments•Expensive and difficult•May create more problems than they solve

Will CFD give the answer?

School of Aeronautical Engineering, Queen’s University Belfast

Cs vs yaw angle

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 10 20 30 40 50 60 70 80 90 100

Yaw angle (deg)

Sid

e fo

rce

coef

fici

ent

(Cs)

QUB 1/50 scale (M-III) Cranfield 1/50 (APT) Turbulence simulation

MIRA 1/5 scale (APT) AEA 1/50 (AEA) Turbulence simulation

Coefficient of side force vs. yaw angle•significant variation between wind tunnel tests•depends on turbulence intensity and scale

School of Aeronautical Engineering, Queen’s University Belfast

Trains with Mark 3 passenger coaches

Class 87 electric loco.+ Mark3 coaches

Prototype for wind tunnel models

High Speed TrainDiesel-electric loco.+ Mark3 coaches

Prototype for CFD model

School of Aeronautical Engineering, Queen’s University Belfast

BMT Fluid Mechanics Wind tunnel (4.8m *2.4m)Atmospheric Boundary Layer simulationClass 87 + Mark 3 coaches (1/30 scale)

School of Aeronautical Engineering, Queen’s University Belfast

Fluent 6Hybrid grid•Unstructured around nose and tail•Relative wind inflow profile specified•Moving ground simulation possible

School of Aeronautical Engineering, Queen’s University Belfast

CFD•Reynolds number 1.5*10^5•Turbulence intensity 3%•Length scale 3m•Yaw angle 60 deg.

Note•Lee side vortex•Separation on ground

School of Aeronautical Engineering, Queen’s University Belfast

QUB wind tunnel experiment•1/50 scale Mark 3 coach & loco.•Flat ground•Turbulence intensity <1%•Reynolds number 1.6*10^5

Cs comparison at V=0.6 m/s

00.1

0.20.3

0.40.50.6

0.70.8

0.91

0 20 40 60 80 100

Yaw angle (degree)

Sid

e F

orc

e C

oef

fici

ent

(Cs)

CFD

Exp

School of Aeronautical Engineering, Queen’s University Belfast

Cs flat ground

0 20 40 60 80 100

Yaw angle (degree)

Sid

e fo

rce

coef

fici

ent

Cs

3% Intensity CFDExp10% Intensity CFD

BMT ABL wind tunnel experiment•1/30 scale Mark 3 coaches & loco.•Reynolds number 2.5*10^5•Turbulence intensity 22% at 3m (full scale)•Turbulence length scale 24m (full scale)

School of Aeronautical Engineering, Queen’s University Belfast

Train on 4m embankment•CFD 3% turbulence•Flow attached to embankment slope•Less sensitive to turbulence?

Cm 4m Embankment

0 20 40 60 80 100

Yaw angle(deg)

Rol

ling

Mom

ent

Coe

ffic

ient

( C

m)

CFD

Exp

School of Aeronautical Engineering, Queen’s University Belfast

CFD shows promise for accurate simulation of •Flow over train on embankment•Forces and moments•Train motion over ground

Problems•Turbulence intensity and length scale important•Ground roughness should be adjusted for correct wind profile

Conclusions

CFD may extend the range of wind tunnel data•Train motion over ground•Unsteady flow, e.g. gust simulation