Flow variability in the vicinity of a busy intersection in Central - - PowerPoint PPT Presentation

Flow variability in the vicinity of a busy intersection in Central London

• A. Balogun, A.S. Tomlin, C. R. Wood, J. Barlow, S. Belcher,
• R. Smalley, J. Lingard, S. Arnold, A. Dobre, A. Robins, D.

Martin and D. Shallcross

Universities of Leeds, Reading, Bristol, Surrey.

Introduction

The aim of the paper: To assess the influence of the roof top background flow on the flow characteristics at the 2 DAPPLE intersection sites. DAPPLE (Dispersion of Air Pollution and Penetration into the Local Environment) 2007 field campaign. An overview of the DAPPLE project and comprehensive description of the measurement site and set-up have been presented in Arnold et al. (2004) for the 2003 campaign and Wood et al. (2009) for the 2007 field

• measurements. Further information is also available at

www.dapple.org.uk. .

Site coordinates & instrumentation layout

R-hand Cartesian coordinate system used (inset). U & V velocity components aligned along Marylebone Rd & Gloucester Pl. 3D ultrasonic sonic anemometers 1 on roof-top reference R 4 @ intersection (~4m and 8m) 1 @ street canyon (~4m )

Measurement Period: May 22-July 4, 2007

Complexity of site geometry

• Marylebone Rd. is widest

canyon so expect strong in street channelling component along MRd.

• Weaker channelling along

Gloucester Pl.

• Buildings different heights

so expect asymmetries.

• Open arc may add to

complexity.

• Site 2 further into canyon

than site 1.

1 2 3

Heights of measurement

Flow influenced by buildings

Zroof = 18.4m H= 21 m

Z ~ 2-3H Z = H Z ~ 0.3H

Thanks to Janet Barlow

11 12 32 21 22

Influence of 15 minute mean roof-top flow direction θref on in street flow θij

Uref : ♦ < 1.2 ms-1, ♦ 1.2 - 3 ms-1, ♦ > 3 ms-1

θij Where I = 1- 3 for site and J = 1- 2 for top & bottom sonics

Influence of 15 minute mean roof-top flow direction θref on in street flow θij

Site 3 shows classic in canyon helical flow pattern.

Uref : ♦ < 1.2 ms-1, ♦ 1.2 - 3 ms-1, ♦ > 3 ms-1

Site 2 - closer to canyon flows than site 1.

• Strong channelling along MRd.

Influence of 15 minute mean roof-top flow direction θref on in street flow θij

Site 3 shows classic in canyon helical flow pattern.

Uref : ♦ < 1.2 ms-1, ♦ 1.2 - 3 ms-1, ♦ > 3 ms-1

Site 2 - closer to canyon flows than site 1.

• Strong channelling along MRd.
• In street recirculation for θref with

significant cross street component

• Weak channelling along GPl (Geometry).

Site 1 bottom sonic

• Combination of complex

flow types leading to high degree of scatter in mean in street flow angle.

• In street channelling for
• nly narrow range of θref.
• Combination of flow

channelling, recirculation, corner vortices and bifurcation all possible for θref oblique.

Uref : ♦ < 1.2 ms-1, ♦ 1.2 - 3 ms-1, ♦ > 3 ms-1.

Site 1 bottom sonic

• Further analysis of complex

regions required to help explain scatter in mean flow angles!

• Wind vector roses & Pdfs

were employed. 46° ≤ θref ≤ 75°

• 134° ≤ θref ≤ -75°

15 minute mean wind roses for oblique reference flows: 46° ≤ θref ≤ 75°

Site 1 shows possible flow bifurcation leading to mean in street angles of anywhere between 0° ≤ θ12 ≤ 120 ° → i.e. large scatter in mean flow angle. Site 2 shows evidence of channelling along MRd, cross canyon recirculation and turning of in street angle – corner vortex?

Roof

Hourly probability density functions (pdf’s) for oblique reference flows.

• Bifurcated flow leads to bi-modal in street

pdf’s at both intersection sites.

• Slight change in θref pdf causes strength of

channelling from different streets to shift at site 1 → scattered mean flow angle.

• Site 2 more consistent – further in canyon.

Recirculation Vortex?

→MRd

 GPl

→MRd

Hourly probability density functions (pdf’s) for oblique reference flows.

• Bifurcated flow leads to bi-modal in street

pdf’s at both intersection sites.

• Changes in θref causes strength of

channelling from different streets to shift at site 1 as the flow is bifurcated→ scattered mean flow angle.

• Site 2 more consistent – further in canyon.

Recirculation Vortex?

→MRd

 GPl

→MRd

Flow visualization and wind tunnel modelling (θref =51°)

Top left Source MRd (south), horizontal light sheet shows intermittent vortex at the south-west corner of the intersection. Bottom right Source at GPl (centre), horizontal light sheet shows vortex at the south-east corner of the intersection. Carpentieri et al, 2009 1 2

15 minute mean wind roses for oblique reference flows: -134° ≤ θref ≤ -105°

Roof

Site 1 shows complex combination of flow patterns leading to huge scatter in mean flow angle:

• bifurcation of channelled

flows along MRd and down GPl

• influence of in canyon

recirculation leading to weak channelling up GPl. Site 2:

• stronger channelling up GPl

than down

• influence of in canyon

recirculation

• backing to the left may

suggest presence of spiralling.

Example hourly pdf’s for

• 134° ≤ θref ≤ -105°

Site 1: Still evidence of channelled modes but far more complex combination of flow patterns than for opposing oblique flows: Bifurcation, recirculation, channelling & influence

• f site geometry.

Site 2: Recirculation dominant.

↓GPl

→MRd Recirculation →MRd

↓GPl

Recirculation

Cross canyon helical vortex Flow bifurcation Corner vortex Fluctuating roof-top wind

Fluctuating background flow leads to changes in relative importance of different flow features → complexity!

Conclusions

• Intersection flows appear to be a complex combination of various flow

patterns (bifurcation, flow recirculation and corner vortices). Results suggests intersection flow pattern is 3D for oblique rooftop flows.

• Site 2 behaviour is closer to typical canyon flows than site 1 even though it

is not deep into the canyon.

• Small changes in background wind direction have pronounced influences
• n the behaviour of intersection flow patterns.
• Consequently, short time-scale variability in background flow direction can

lead to highly scattered in-street mean flow angles.

• Averaged flow angles hide the true multi-modal features of the flow!
• Geometric features at the intersection corners also significantly influence

the variability of flow behaviour at both sites. Modelling effort will need to incorporate this for improved accuracy.

Acknowledgements

We acknowledge EPSRC and the UK Home Office for DAPPLE funding. We also thank DAPPLE colleagues, staff at WCCH, Transport for London and the School of Earth and Environment at the University of Leeds.

Cross canyon helical vortex Flow bifurcation Corner vortex Fluctuating roof-top wind

Fluctuating background flow leads to changes in relative importance of different flow features → complexity!

Thank you for your attention