Numerical and experimental modelling of pollutant dispersion in a street canyon

https://doi.org/10.1016/S0167-6105(01)00215-XGet rights and content

Abstract

The pollutant dispersion in a two-dimensional street canyon is studied in this project. The principal parameter investigated is the height of the downstream building. The pollutant source is situated in the middle of the street. The investigation is performed in two ways. Experiments have been carried out in the L-2B wind tunnel at von Karman Institute and numerical simulations have been done with the CFD software Fluent 5.2. The concentration measurements have been performed by means of light scattering technique and the velocity field has been measured with particle image velocimetry. In the numerical simulations, a preliminary study about the backward-facing step has been performed in order to select the best turbulence model in Fluent for these complex flows characterized by separation, stagnation, recirculation, reattachment, etc. The best model appeared to be the realizable kε model with the two-layer zonal approach to the wall, which predicts the reattachment length after the step with <1% error in comparison with the value obtained from direct numerical simulation by Le, Moin and Kim (Direct numerical simulation of turbulent flow over a backward-facing step, Report No. TF-58, 1996). This model has been applied in the street canyon simulations. Comparison with the experimental results has been made.

Besides the height of the downstream building, the influence of a third building situated upstream of the street canyon in the flow and dispersion inside the street has been investigated.

Introduction

One of the principal sources of contamination (nitrogen oxides and hydrocarbons) in the cities are the transport vehicles and a worsening of the situation may be expected in view of the continuous increase in traffic. Along the same city streets (urban canyons) people are walking unprotected, and windows are open along the side buildings with people sitting inside often for long times. Therefore, it is of major importance to find out how these pollutants distribute in the streets and if, modifying some parameters, the pollution at pedestrian level and along the wall can be decreased. Configurations like this have been investigated by Meroney et al. [2], Leitl and Meroney [5], Murakami [10], Zhang et al. [7], Hwang et al. [8], Tsuchiya et al. [9] and many others.

A limitation of direct field measurements of atmospheric phenomena [1] is that all possible governing parameters are simultaneously operative and it is not simple to determine which are important, which are secondary or which are insignificant. These parameters are: building geometry (height, width, roof shape), street dimensions (breadth, width), thermal stratification (solar insulation and orientation, building and street thermal capacitance), vehicular movement (size, number, frequency), plume buoyancy etc., and they are all intertwined.

Wind tunnel experiments provide an opportunity to examine the effects of various parameters individually or in combination. In this case, the street canyon has been modelled as two blocks of wood which extend over the whole width of the test section. The wind direction is perpendicular to the street to preserve two dimensionality.

The pollutant concentration has been measured using smoke as a tracer and recorded by a video camera light scattering technique. The images have been analysed by a digital image processing system. This approach samples the whole two-dimensional flow field providing an instantaneous image of the distribution of pollutant. The mean concentration field has been obtained after the processing of the images. In order to examine the transport mechanism in the canyon, the velocity field has been measured with particle image velocimetry (PIV). Simultaneously, the possibility of predicting the pollutant concentration numerically has been investigated. The flow field around “bluff bodies” such as buildings with sharp edges and corners is very complex to compute by a CFD code. In this research project, a systematic study about the numerical simulations has been done using the different options available in the FLUENT 5.2 code. The best turbulence model for separated and recirculating flows has been selected after a preliminary study on the wake of a backward-facing step and this model has been applied to the street canyon simulations.

The results presented in this paper are mostly concerned with the influence of the relative height of the upstream and downstream buildings in the pollutant dispersion inside these canyons for an isolated street (open country) and for the case in which it is preceded by an upstream building and road (non-isolated street canyon) under a wind coming normal to the main streets.

Section snippets

Street canyon model

A simple geometry represented by two blocks of wood spanning the width of the wind tunnel has been used to model the street canyon in order to exclude three-dimensional effects. The blocks have been painted in black to avoid the reflections of light from the laser sheet. The dimensions of the blocks were 3cm×3cm×34.8cm. The height of the downstream building, as stated before, was the main variable. Three different values were used: 3,4 and 5cm. A sketch of the model with buildings of equal

Experimental set-up

The experiments have been conducted in the L-2B wind tunnel at von Karman Institute (Fig. 2).

This facility is a low-speed, open-circuit wind tunnel of suction type. It incorporates an air inlet, fitted with honeycombs, meshes and a two-dimensional contraction. The test section is a square of 0.35m size and a length of 2m. The maximum achievable velocity is 35m/s. The model of the street canyon is placed in the middle of the test section and spans its full width.

The free stream velocity for the

Velocity measurements: PIV

The free stream velocity and the boundary layer profile has been measured with a hot wire. For the velocity field around and inside the street canyon PIV has been used. The principal advantages of PIV are the capability to determine instantaneous velocity field and its quantification over the entire plane.

To solve the directional ambiguity associated to PIV two different and successive images were recorded and used for the analysis.

The acquisition of the images was done with INSIGHT software of

Open country configuration

In Fig. 7, Fig. 9, Fig. 11 the mean velocity field from PIV is represented for the open country configuration. In the three figures it is possible to see how the flow is separated and accelerated when impacting against the upstream building. A recirculating region is formed above the street when the buildings have the same height (Fig. 7). This region is broken by the downstream building when increasing its size, since the flow over and after the canyon is not allowed to go backwards.

A strong

Preliminary investigation

A preliminary investigation has been performed to select the turbulence model in the commercial software Fluent 5.2 more suitable for this kind of complex flows, which are defined by stagnation, separation, recirculation, reattachment, etc. Therefore, the flow over the backward-facing step has been computed using different models of turbulence. The results have been compared with the solution from direct numerical simulation of turbulent flow over a backward-facing step by Le, Moin and Kim [4].

Conclusions

The pollutant dispersion in a street canyon has been investigated by means of experiments in the L-2B wind tunnel at von Karman Institute (VKI) and numerical simulations with the CFD code FLUENT (version 5.2). The principal parameter to investigate has been the height of the downstream building.

Two different sets of configurations have been defined: open country and non-isolated street canyon. In the open country cases, only the street canyon is studied without any other buildings around. In

References (10)

There are more references available in the full text version of this article.

Cited by (105)

  • Revisit of prevailing practice guidelines and investigation of topographical treatment techniques in CFD-Based air ventilation assessments

    2020, Building and Environment
    Citation Excerpt :

    CFD can provide complete data for the flow and concentration field within the computational domain and can be performed at both reduced and full scales. In addition, CFD models enable efficient parametric and sensitivity analysis [32–43] and, most importantly, can provide intuitive visualization to aid the understanding of complex ongoing physical processes within an urban canopy. In the past decade, extensive AVA studies have been conducted using CFD with reference to the guidelines stated in the AVA TC.

  • Exploration of sustainable building morphologies for effective passive pollutant dispersion within compact urban environments

    2019, Building and Environment
    Citation Excerpt :

    CFD can provide complete data for the flow and concentration field within the computational domain and can be performed at both reduced and full scale. Additionally, CFD models are also efficient for parametric and sensitivity analysis [38−48], and most importantly, the tool can provide good visualization to aid in understanding of the complex on-going physical processes within an urban canopy. The recent review paper by Tominaga and Stathopoulos [49] summarizes several aspects of CFD modelling for pollutant dispersion problems.

  • Influence of photochemical processes on traffic-related airborne pollutants in urban street canyon

    2016, Journal of Atmospheric and Solar-Terrestrial Physics
    Citation Excerpt :

    These and other studies have identified many factors that affect the flow and dispersion of pollutants in street canyons, including ambient wind speed and direction, building geometry, and thermal stratification (Berger et al., 1997; Hu et al., 2012; Kang et al., 2008; Zelinger et al., 2004). Mathematical models employing numerical calculations or methods based on physical modelling (Civis et al., 2001; Garcia Sagrado et al., 2002; Zelinger et al., 2009, 2006) are usually applied to overcome the lack of measurements obtained from the atmosphere. To validate the numerical models (Boehm and Aylor, 2005; Di Sabatino et al., 2007; Khan and Abbasi, 2000; Pospisil et al., 2004), the use of wind tunnel experiments (Baik et al., 2000; Kastner-Klein and Plate, 1999; Zelinger et al., 1999) is usually strongly recommended.

View all citing articles on Scopus
1

Colour figures can be obtained upon request.

View full text