CFD modelling of tree-atmosphere interaction and ... · PDF fileCFD modelling of tree-atmosphere interaction and ... CFD modelling of tree-atmosphere interaction and assessment of

CFD modelling of tree-atmosphere interaction andassessment of air quality in street canyonsSabatino, Di, S.; Buccolieri, R.; Gromke, C.B.

Published in:Proceedings of the CLIMAQS Workshop ‘Local Air Quality and its Interactions with Vegetation’, January 21-22,2010, Antwerp, Belgium

Published: 01/01/2010

Document VersionPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

Citation for published version (APA):Sabatino, Di, S., Buccolieri, R., & Gromke, C. (2010). CFD modelling of tree-atmosphere interaction andassessment of air quality in street canyons. In Proceedings of the CLIMAQS Workshop ‘Local Air Quality and itsInteractions with Vegetation’, January 21-22, 2010, Antwerp, Belgium (pp. 1-50)

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Download date: 24. May. 2018

https://research.tue.nl/en/publications/cfd-modelling-of-treeatmosphere-interaction-and-assessment-of-air-quality-in-street-canyons(eccdc005-ce10-48c5-afb7-1e0a6f58a00e).html

CFD modelling of tree-atmosphere interaction and CFD modelling of tree-atmosphere interaction and assessment of air quality in street canyonsassessment of air quality in street canyons

SILVANA DI SABATINO Dipartimento di Scienza dei Materiali - University of Salento (ITALY)

[email protected]

RICCARDO BUCCOLIERI Dipartimento di Informatica - Università “Cà Foscari” di Venezia (ITALY)

Dipartimento di Scienza dei Materiali - University of Salento (ITALY)[email protected]

CHRISTOF GROMKE WSL Institute for Snow and Avalanche Research SLF (Switzerland)

Institute for Hydromechanics - University of Karlsruhe (GERMANY)[email protected]

Antwerp, Belgium, January 21-22, 2010

Local Air Quality and its Interactions with Vegetation

UNIVERSITY OF SALENTO (ITALY)

Local Air Quality and its Interactions with Vegetation, Antwerp, Belgium, January 21-22, 2010 C

FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Introduction - Background ideas / urban areas (buildings, trees ..)

Previous work - Single tree - Line of vegetation - Trees in street canyons

CFD simulations (RANS mainly)/validation - Aerodynamic effects of trees in street canyons - Aerodynamic effects of tree in complex geometries (IDEALISED)

– Application to a real case scenario - Bari city (Italy)

Conclusions and future perspective

OutlineOutline


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Urban surface is source of drag, heat and pollution Street (canyon)

scale

Neighbourhood scale

City scale

Flow and dispersion modelling is currently possible at many levels of sophistication depending on applications

IntroductionIntroductionScale definitions


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

STREET CANYONSTREET CANYON

aspect ratioaspect ratio, W/H

city basic geometry unit

geometries which affect flow and turbulence fields

where the people and (the emissions) arewhere trees can be planteddirect CFD/LES is practicableoperational modeling is typically based on a more idealized recalculating vortex driven by a shear layertraffic pollutants released near the ground need to be “effectively” dispersed to maintain “adequate” air quality

Street canyon

IntroductionIntroductionUrban street canyons


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

3-D flow field inside a street canyon with aspect ratio W/H ≈ 1

Oke, T.R., 1987. Boundary Layer Climates, 2nd Edition. Methuen, UK..

• approaching flow perpendicular to street axis• long street canyons (L/H > 7)• two dominating vortex structures: canyon vortex, corner vortex• superposition of vortex structures

IntroductionIntroductionQualitative Flow features


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

IntroductionIntroductionExample of Urban street canyons

Street canyon without trees (but with people…!)

Street canyon with one-row trees Street canyon with two-rows trees


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Impact of trees in urban areas on pollutant dispersion not widely considered

Both experimental and numerical investigations are present in the literature

Some of the tree effects on flow and dispersion have been considered individually in previous works, such as deposition, filtration, blockage etc.

Still far from a comprehensive understanding of the overall role plaid by vegetation on urban air quality

One of the most extensive review is given by Litschke and Kuttler (2008), who reported on several field studies as well as numerical and physical modelling of filtration performance of plants with respect to atmospheric dust. The main goal of this review was to assess whether a reduction in particle concentration in the air could be accomplished by existing vegetation or targeted planting.

Where are we?Where are we?

Litschke, T and Kuttler, W., 2008. On the reduction of urban particle concentration by vegetation – a review. Meteorologische Zeitschrift 17, 229-240.


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Some tree effectsSome tree effectsLitschke and Kuttler (2008)

The deposition of particles on plant surfaces is influenced by a variety of factors:

diameter and shape of the particlesplanting configuration meteorological parameters (air relative humidity, wind speed and turbulence)

Deposition velocity (vd) is used as a measure of filtration performance

Fp: mass particle flow rate Cp: leaf surface and atmospheric particle concentration


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

In urban areas, values seems not to exceed vd < 1 cm s−1.

At these deposition velocities, the improvement in air quality as a result of filtration by plants is about 1%; this implies that in order to compensate for local particulate emissions by road traffic, very large areas of vegetation would be needed.

Some tree effects: FiltrationSome tree effects: FiltrationLitschke and Kuttler (2008)


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

obstacles to airflow (air mass exchange reduced)

Some observationsSome observationsLitschke and Kuttler (2008)

particle deposition on plant surfaces pollutant concentration reduced

pollutant concentration increased

New measurement method by Freer-Smith et al. (2005)

Filtration performance, in some cases, IS an order of magnitude higher than the values published to date

However, the in-situ measurements using the method proposed by Freer-Smith et al. (2005) will still need to be confirmed by further measurements

Deposition velocities of PM10, PM2 und PM1 on the foliage of Pinus sylvestris, X Cupressoparis leylandii, Acer campestre,Populus deltoides X trichocarpa ‘Beaupré’ and Sorbus aria (modified after FREER-SMITH et al., 2005).


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Tree Canopy ModellingTree Canopy ModellingMochida et al. (2008)

Extensive review on past tree canopy models based on the k-ε model in which extra terms were added in the transport equations to simulate the aerodynamic effects of trees, both on velocity decrease and increase in the turbulence and energy dissipation rate. Canopy models classified based on the different forms of extra terms adopted:

5 parameters in the extra terms

Cpε1, Cpε2: two model coefficients

3 extra terms:

-Fi: effect of trees on velocity decrease+Fk and +Fε: effects of trees on the amount of increase in turbulence and energy dissipation rate, respectively.

Mochida, A., Tabata, Y., Iwata, T. and Yoshino, H., 2008. Examining tree canopy models for CFD prediction of wind environment at pedestrian level. Journal of Wind Engineering and Industrial Aerodynamics 96, 1667-1677.


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

FOCUS

1. Flow and dispersion around a single tree (field and numerical results)

1. Flow and dispersion around a line of vegetation (field and numerical results)

1. Flow and dispersion in street canyon with tree planting (field, wind tunnel and numerical results)

Example of recent results Example of recent results (not comprehensive!)(not comprehensive!)


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Field and laboratory studies

Several field campaigns, wind tunnel experiments and numerical simulations were performed to investigate the effects of trees on outdoor airflow distribution, turbulent diffusion of air pollutants and thermal environment in street canyons.

Field experiments: in Metropolitan Tokyo (Narita et al., 2008) in the central part of Sendai city, Japan (Kikuchi et al, 2009).

Wind tunnel experiments: CODASC, 2008 (University of Kalshrue)

Numerical simulations: e.g. Balczó et al., 2009 Gromke et al., 2008; Buccolieri et al., 2009

Literature review

Our work


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Field experiments (Narita et al., 2008)

Street canyon with trees (G-street)

Empty street canyon (non G-street)

Narita, K., Sugawara, H. and Honjo, T., 2008. Effects of roadside trees on thermal environment within street canyon. Geographical Report of Tokyo Metropolitan University 43.

Metropolitan Tokyo

Field studies


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

The downward short wave radiation over the street with tree crown is about one-third of that without tree crown in midday. It shows clear shielding effects on solar radiation by “green tunnel”.

On the contrary, the downward long wave radiation under tree crown is about 20 – 30 Wm-2 larger than that of without tree crown throughout the day.

Field experiments (Narita et al., 2008) with treesempty

Field studies


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

The air temperature and absolute humidity differences between the two streets are smaller than expected

There are few reports which implied that the increase of humidity in dense green area has an adverse influence on human comfort. In this case, however, there is no evidence for such a negative effect by the tree crown.

The wind speed difference is not so large

A possible negative factor of roadside trees for sensational climate is to weaken the wind speed of pedestrian level.


with treesempty

Field studies


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

During the day, because of the shading effect of the tree crown, the area-averaged road surface temperature in “G-street” is about 8 degrees lower than that in “non-G-street”.

During the night road surface temperature under the tree crown is one or one and half degrees higher than that of without tree crown. That is because the radiative cooling is diminished by the tree crown in the “G-street”.

In both streets, these temperature differences are always positive, that means road surface is heat source for the air in the street canyon even though under the tree crown.

Field experiments (Narita et al., 2008) with treesempty

Field studies


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Though roadside trees are expected as an important mitigation method for heat island phenomena, the difference of air temperature is negligible throughout the day comparing to the parallel street without tree crown.

The increase of humidity and the decrease of wind speed are not so discernible. As for the radiation field, however, there is clear shielding effects for solar radiation and downward longwave radiation, consequently the road surface temperature.

That means the main reason to feel comfortable under tree crown in thermal sensation is not the air temperature but the radiation effects.


Field studies


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

•Gross G. (1987): •influence tree planting (two rows arranged sidewise of the street next to the building walls) •k-ε turbulence closure scheme and modelling of the tree crowns by porous bodies•decelerated flow velocities near the building walls and increased pollutant concentrations inside the street canyon

•In a similar arrangement, Ries, K. and J. Eichhorn (2001) found local increases of the pollution concentration at the leeward wall accompanied by reduced flow velocities due to trees.

2D models applied (do not account for the highly 3D flow fields present in real urban street canyons of finite length)

Gross, G., 1987. A numerical study of the air flow within and around a single tree. Boundary-Layer Meteorology 40, 311-327.

Ries, K. and Eichhorn, J., 2001. Simulation of effects of vegetation on the dispersion of pollutants in street canyons. Meteorologische Zeitschrift 10, 229-233.

Numerical investigations


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

WIND TUNNEL INVESTIGATIONSGromke, C. and Ruck, B., 2009. On the impact of trees on dispersion processes of traffic emissions in street canyons. Boundary-Layer Meteorology 131, 19-34.

CODASC, 2008. Concentration Data of Street Canyon, internet database, http://www.codasc.de. Institute for Hydromechanics - University of Karlsruhe (GERMANY)

Flow and concentration fields in urban street canyons of different aspect ratios with various avenue-like tree planting configurations

Tree planting characteristics: influence of crown shape, diameter, height, porosity and planting density

FLOW: air exchange and entrainment conditions considerably modified, resulting in lower flow velocities and in overall larger pollutant charges inside the canyon.

DISPERSION: increases in pollutant concentrations at the leeward and decreases at the windward

street canyon with miscellaneous tree arrangements

Model validation – but not only!


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Example of a typical CFD simulation setup

• commercial CFD-Code• RANS-Equations• turbulence closure schemes

- RSM at least!• second order discretization schemes• grid: hexahedral elements

- ~ 400,000- δx=0.05H, δy=0.25H, δz=0.05H

- expansion rate <1.3 • turbulent Schmidt number Sct = 0.7

ν=ydiffusivitturbulent

itycosvisturbulentD

Sct

tt

uH=4.7 m/s: undisturbed wind speed at the building height H

α=0.30: power law exponent

=0.52 m/s: friction velocity

κ=0.40: von Kàrmàn constant

Cμ = 0.09

α

=

Hz

uzuH

)()

δz(

Cu

kμ

−= ∗ 12

)δz(

κzuε −= ∗ 1

3

INLET

30H8H

8H

INFLOW

OU

TFLOW

SYMMETRY

WIND

lQHucc

T

refm=+

cm measured concentration

uref reference velocity

H building height

QT/l strength of line source

Dimensionless concentrations c+

CFD modelling CFD modelling Objectives: Validation studies / speculative approach

*u


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

CFD modelling of porous tree crowns

A cell zone is defined in which the porous media model is applied and the pressure loss in the flow is determined

The porous media model adds a momentum sink in the governing momentum equations:

This momentum sink contributes to the pressure gradient in the porous cell, creating a pressure drop that is proportional to the fluid velocity (or velocity squared) in the cell.

The standard conservation equations for turbulence quantities is solved in the porous medium. Turbulence in the medium is treated as though the solid medium has no effect on the turbulence generation or dissipation rates.

viscous loss term + inertial loss term

Si: source term for the i-th (x, y, or z) momentum equation : magnitude of the velocity D and C: prescribed matricesv

permeable zone with the same loss coefficient λ as in wind tunnel experiments

LOOSELY FILLED: λ = 80 m-1

DENSELY FILLED: λ = 200 m-1


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

single block: impermeable

single block: permeable (λ=250m-1)

Empty street canyon

Street canyon with tree planting along the center axis

Gromke, C., Buccolieri, R., Di Sabatino, S. and Ruck, B., 2008. Dispersion study in a street canyon with tree planting by means of wind tunnel and numerical investigations - Evaluation of CFD data with experimental data. Atmospheric Environment 42, 8640-8650.

VALIDATION STUDIES (W/H=1)


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

-0.4 -0.2 0 0.2 0.4x/H

LDV Measurement

0

0.2

0.4

0.6

0.8

1

1.2

z/H

-0.4 -0.2 0 0.2 0.4x/H

k-epsilon TM

0

0.2

0.4

0.6

0.8

1

1.2z/

H

-0.4 -0.2 0 0.2 0.4x/H

RSM TM

0

0.2

0.4

0.6

0.8

1

1.2

z/H

w velocities – middle of the canyon

-0.4 -0.2 0 0.2 0.4x/H

k-epsilon TM

0

0.2

0.4

0.6

0.8

1

1.2

z/H

-0.4 -0.2 0 0.2 0.4x/H

RSM TM

0

0.2

0.4

0.6

0.8

1

1.2

z/H

-0.4 -0.2 0 0.2 0.4x/H

LDV Measurement

0

0.2

0.4

0.6

0.8

1

1.2

z/H

empty street canyon permeable crown

•slightly smaller flow velocities in the upward •significantly lower velocities in the downward moving part•volume flow crossing the horizontal plane at z/H = 0.7 is reduced (-36 %)

canyon vortex

Air volume rotating in canyon vortex is reduced in the presence of tree plantings


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

• Is wind direction important? Competition with aspect ratio…

• Approaching flow perpendicular and inclined by 45° to street axis

Empty street canyon - W/H=2

Street canyon with tree planting

CFD modellingCFD modelling

Buccolieri, R., Gromke, C., Di Sabatino, S. and Ruck, B., 2009. Aerodynamic effects of trees on pollutant concentration in street canyons. Science of the Total Environment 407, 5247-5256.


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

WIND

Wall A shows larger concentrations than windward wall B: dominating vortex-like structure around the street canyon centre

Decreasing concentrations towards the street ends: enhanced natural ventilation at the street canyon ends, where air exchange is not only provided by vertical exchange with the above roof flow but also with entering flow laterally

x/H

z/H

-1 -0.6 -0.2 0.2 0.6 10

0.2

0.4

0.6

0.8

1

1.2

0.1

0.2

0.3

0.4

-1 -0.5 0 0.5 1-6

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

y /H

x/H

WIND

z=0.5Hy=1.25H

EMPTY STREET CANYON

W/H=2, wind direction: perpendicular

CFD FLOW INSPECTIONCFD FLOW INSPECTION


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Wall A Wall B

CFD concentration pattern qualitatively similar to that obtained in the wind tunnel

Slight underestimation of the measured concentrations in proximity of wall A

calculated concentrations relative deviations [%] in respect of measurements

street canyon model – wind tunnel

street canyon model – CFD

VALIDATIONVALIDATIONEMPTY STREET CANYON



FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Wall A Wall B

relative deviations [%] in respect of tree-less street canyon

Increases in concentrations in proximity of wall A and decreases near wall B

Maximum concentrations at pedestrian level in proximity of wall A

Differently to the tree-free street canyon case, less direct transport of pollutants from wall A to wall B occurs

WT CONCENTRATIONSWT CONCENTRATIONS

Loosely filled crown Loosely filled crown ((λ = 80 m-1 , PVol = 97.5 %)



FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Increases in concentrations in proximity of wall A and decreases near wall B

The pollutants are advected towards the leeward wall A, but, since the circulating fluid mass is reduced in the presence of tree planting, the concentration in the uprising part of the canyon vortex in front of wall A is larger

-1 -0.6 -0.2 0.2 0.6 10

0.2

0.4

0.6

0.8

1

1.2

x/H

z/H

0.1

0.2

0.3

0.4

-1 -0.5 0 0.5 1-6

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

y/H

x/H

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

x/H

z/H

-1 -0.6 -0.2 0.2 0.6 10

0.2

0.4

0.6

0.8

1

1.2

0.1

0.2

0.3

0.4

-1 -0.5 0 0.5 1-6

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1y/H

x/H

Differently to the tree-free street canyon case, less direct transport of pollutants from wall A to wall B occurs

Most of the uprising canyon vortex is intruded into the flow above the roof level. Here, it is diluted before partially re-entrained into the canyon. As a consequence, lower traffic exhaust concentrations are present in proximity of wall B

WIND

WIND z=0.5Hy=1.25H

WITH TREES


CFD FLOW INSPECTIONCFD FLOW INSPECTION


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns


CFD simulations were successful in predicting an increase in concentrations in proximity of wall A and a decrease near wall B and the relative deviations in respect of tree-less street canyon

As in the tree-free case, it slightly underestimated experimental data

Wall A Wall B


street canyon model – CFD

Loosely filled crown Loosely filled crown ((λ = 80 m-1 , PVol = 97.5 %)


VALIDATIONVALIDATION


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

In comparison to the street canyon with the tree planting of loosely filled crown, no essentialchanges are found both in wind tunnel experiments and CFD simulations

The degree of crown porosity is of minor relevance for flow and dispersion processes inside the street canyon as the tree planting is arranged in a sheltered position with wind speeds being very small.

WT CONCENTRATIONSWT CONCENTRATIONS

relative deviations [%] in respect of tree-less street canyon

CFD - WT CONCENTRATIONSCFD - WT CONCENTRATIONS


Densely filled crown Densely filled crown ((λ = 200 m-1 , PVol = 96 %)

W/H=2, wind direction: perpendicularVALIDATIONVALIDATION


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

NMSE R FAC2 FB

tree-less street canyon 0.06 0.96 0.97 0.15

loosely filled crown (PVol = 97.5 %) 0.13 0.98 1.00 0.21

densely filled crown (PVol = 96 %) 0.09 0.99 1.00 0.14

Statistical analysis (Chang, C. and S. Hanna, 2004):

•normalized mean square error (NMSE ≤ 4)•correlation coefficient (R)•fraction of predictions within a factor of two of observations (FAC2 ≥ 0.5)•fractional bias (−0.3 ≤ FB ≤ 0.3)

Comparison Wind Tunnel Measurements - Numerical Computations CFD simulations are in general good agreement with wind tunnel experiments

Pollutant concentrations in proximity of the leeward wall are slightly underestimated in the numerical

simulations, while near the windward wall both slightly over- and underestimations are present

W/H=2, wind direction: perpendicularVALIDATIONVALIDATION


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Relative deviation in wind tunnel

concentration

H/W=1 –single tree row vs empty

H/W=0.5 –two tree rows vs empty

leeward +71% +42%windward -35% -32%

Concentration fields within street canyon depend on both street canyon

aspect ratio and tree planting configuration

Double tree rows is preferable to one row in the middle of the canyon

Some observationsSome observationsWind direction: perpendicular


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

W/H=2, wind direction: 45°

Pollutant concentrations are larger than at wall B. Concentration increases from the centre to the street ends at both walls are found.

In the wind tunnel experiments, at the beginning of wall A large concentrations are found. This phenomenon is y only partially reproduced in the CFD simulations.

Overall CFD concentrations are similar to those obtained in the wind tunnel, even if there is some underestimation of the measured concentrations at wall A.


Wall A Wall B



FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Wall A Wall B


CFD FLOWCFD FLOW


Lower concentrations at both walls at the upstream entry are due to enhanced ventilation caused by the superposition of the canyon vortex and the corner eddy.

The increasing pollutant concentrations towards the downstream end of the street clearly indicate that the flow along the street axis becomes a dominant pollutant transport mechanism.

This tendency is due to the helical flow characteristic of the canyon vortex. Moreover, the clockwise rotating helical motion determine the vertical concentration distributions on both walls.


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Increases in concentrations at both wall.

Overall CFD concentrations are similar to those obtained in the wind tunnel, even if there is an underestimation of the measured concentrations at wall A, especially close to the upstream entry.


Wall A Wall B



Densely filled crown Densely filled crown ((λ = 200 m-1 , PVol = 96 %)


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Concentration patterns are due to the predominant parallel flow component.

In particular, at the upstream entry of wall A the corner eddy found in the tree-free case does not occur anymore, due to the presence of trees which behave as obstacles

The helical flow vortex is also broken and, as a consequence, a wind flowing parallel to the walls is evident. However, from the figure it can be noted that wind velocities are slower than those found in the previous case. As the result of this, the pollutants released from the traffic are larger.

CFD FLOWCFD FLOW

Wall A Wall B




FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

For the 45° inclined case there is an increase in concentrations both at leeward

and at the windward!

Concentration fields within street canyons depend on approaching wind direction

rather than on tree planting arrangement

Wind tunnel concentrations

Perpendicular 45° inclinedleeward windward leeward windward

Tree-free(CFD results)

14.8(12.1)

5.2(5.4)

9.8(7.4)

0.9(0.9)

Tree vs tree-free(CFD results)

~ +42%(~ +40%)

~ -32%(~ -24%)

~ +89% (~ +72%)

~ +211% (~ +344%)

Increase Increase IncreaseDecrease

W/H=2, Effect of wind direction


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

For the perpendicular approaching flow the increase in concentration due to tree planting is larger for W/H = 1 case

For the 45° inclined flow the increase due to tree planting is larger for W/H = 2 case.

Wind tunnel concentrations

Perpendicular 45° inclinedW/H = 1 W/H = 2 W/H = 1 W/H = 2

Tree vs tree-free

(CFD results)

Larger ~ +48%

Smaller ~ +23%

(~ +24%)

Smaller ~ +65%

Larger ~ +99%

(~ +101%)

Effect of aspect ratio vs. wind direction

Comparison of overall (at both walls) concentration increase due to tree planting

Volume occupied by tree crowns

W/H=1: ~33%

W/H=2: ~ 28%


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

REAL SCENARIOSREAL SCENARIOSAerodynamic effects of trees in Bari (Italy)

1

2

3

4

4 street canyons and 1 junction

Hmax~46m, Hmean~24m

Four tree rows Avenue-like tree planting of high stand densities, i.e. with interfering neighbouring tree crowns.

Bari (ITALY)


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

The variability of the wind speed, the presence of background and other variables not controlled in the field experiments limit the comparison between measured and simulated data to qualitative analysis of the concentrations. CFD simulations aim at providing an example of how numerical tools can support city planning requirements Computational cells: three millions and a half (cell dimensions δxmin = δymin = 1m, δzmin = 0.3m until the height of 4m). 4 days simulation time with 2 processors

- street canyon 1-2: W/H ~ 0.5- street canyon 1-3: W/H ~ 1.8- street canyon 2-4: W/H ~ 1.8- street canyon 3-4: W/H ~ 0.6

Wind dir.: 5°


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

•15 December 2005

•Wind dir.: West

•Larger conc.: 33μg/m3

•26 December 2005

•Wind dir.: South

•Smaller conc. 18μg/m3

Measurements at monitoring station (~3m)

c+ ~ 20

c+ ~ 16

CFD simulations

lQHucc

T

refm=+cm modeled concentration

uref reference velocity

H tallest building height

QT/l strength of line source

33/18 ~ 1.8 (MEAS.)

20/16 ~ 1.3 (SIM.)


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Flow pattern within the canyon can be thought as an interaction of several approaching winds leading to an overall increase of concentration at both sides of the canyon with trees as found for the 45° inclined case for isolated canyons.

•Larger velocities

•3 times smaller concentrations at monitoring position without trees

Wind dir.:5°

Real case (with trees)

Case without trees

c+ c+

U (m/s)c+


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Also in this case, flow pattern can be thought as an interaction of several approaching winds leading to an overall increase of concentration at both sides of the canyon with trees as found for the 45° inclined case for isolated canyons.

•Slightly larger velocities (channelling along tree spaces transports more pollutant away from monitoring position)

•1.3 times larger concentrations at monitoring position without trees

Real case (with trees)

Case without trees

c+ c+

U (m/s)c+


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Practical relevance of results Practical relevance of results

Concentrations MEASUREMENTS SIMULATIONS

West South West/South West South West/South

Tree 33μg/m3 18μg/m3 ~1.8 20 16 ~1.3Tree-free N/A N/A N/A 6 21 ~0.3

Simulations show that it has been crucial to consider the effect of trees on pollutant dispersion to explain qualitative difference between monitoring stations

Without trees the situation is reversed!


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Work in ProgressWork in ProgressLarge Eddy Simulation – Transient Solution

Simulations were performed by Salim Mohamed Salim, Division of Environment, Faculty of Engineering, University of Nottingham (Malaysia Campus), Semenyih, 43500, Selangor Darul Ehsan, Malaysia

Preliminary LES work

• Started with RANS mean flow steady solution-Standard k-ε

• Imposed synthetic turbulence & continued with Large Eddy Simulation

- Transient solution• subgrid-scale model

- Dynamic Smagorinsky-Lilly• bounded central differencing for momentum, second

order for species & energy.• grid: hexahedral elements

- ~ 550,000- δx=0.1H, δy=0.1H, δz=0.1H

• Flow-through-time (residence time),- Tf = L/U = 41 X 0.12 m / 6.2ms-1 = 0.8s

- L: length of domain in x-direction- U: undisturbed velocity

• timestep size , Δt = 0.005• initial number of timesteps:4000 (= 20sec= 25Tf )

- statistically steady state achieved• additional 4000 timesteps, to average the instantaneous velocity

(bringing total to 40sec)• Computational cost:

- Intel® Xeon® Processor W3520 (8M Cache, 2.66 GHz, 4.80 GT/s Intel® QPI)

- Parallel Processing on 4 Cores.- 18 iterations per time-step for convergence at 2.5 sec/iteration

45sec/timestep 100hrs of computation.


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Work in ProgressWork in ProgressLarge Eddy Simulation – Transient Solution

t = 20s

t = 35s

t = 30s

t = 25s

t = 40s

Time-averaged at 40s

Wall A Wall B

t = 20s t = 30s Time-averaged at 40s

• Concentration levels on the windward (Wall B) and leeward (Wall A) walls.

• Velocity contours showing instantaneous (t=20s and 30s) and mean values.


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

ConclusionsConclusions

Trees in urban street canyons have important aerodynamics effects(aspect ratios and wind direction are among the most important ones!) They have somehow been quantified using wind tunnel controlled experiments. Real conditions may be different. BULK effects are probably understood individually but not in combination (especially in real scenarios)… multiple canyons, neighbourhood scale

RANS CFD simulations/analyses for concentration predictions in street canyons are currently feasible with a proper turbulence closure but most probably LES is more adequate to take into account non-stationary processes

We still need to account for the effect of buoyancy (Radiation Sheltering effect but buildings release heat in. Trapping effect. Warm air in the bottom part of the canyon


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

Future PerspectiveFuture Perspective

More good experiments are needed to further investigate the complex atmosphere-tree interaction in urban areas of both fundamental and applied nature are necessary (velocity and turbulence time scales, momentum and scalars transfer processes …). Integrated (perhaps holistic approach) to understand/predict the urban environment. Trees and vegetation is part of the system

Only by a synergic effort between methodologies and scientific expertise will be possible to tackle the new challenges associated with the emerging properties/difficulties/prediction of the urban environment in a changing climate!


FD m

odel

ling

of tr

ee-a

tmos

pher

e in

tera

ctio

n an

d as

sess

men

t of a

ir q

ualit

y in

stre

et c

anyo

ns

THANK YOUR

FOR

YOUR ATTENTION

CFD modelling of tree-atmosphere interaction and ... · PDF fileCFD modelling of tree-atmosphere interaction and ... CFD modelling of tree-atmosphere interaction and assessment of

Documents