Pollutant transport between a street canyon and a 3D urban array as a function of wind direction and roof height non-uniformity

Measurement quantities

First, the nomenclature of qunatitites is introduced. The lower (e.g. $u$ ) and upper (e.g. $U$ ) case letters stand for the instantaneous value and the mean value respectively, while the primed letter (e.g. $u'$ } and the overlined bar (e.g. ${\overline {u}}$ ) stand for the turbulent fluctuations and the temporal averaging respectively. Since the two instantaneous velocity components (longitudinal, $u$ , and vertical, $w$ , for the top openings and longitudinal, $u$ , and lateral, $v$ , for the lateral openings) and the pollutant concentration ( $c$ ) were measured simultaneously at each given point, the following quantities normalised to the reference flow velocity $U_{ref}$ (which was the freestream velocity of the wind tunnel) were calculated from 120-s time series:

The dimensionless mean longitudinal ( $U/U_{ref}$ ), vertical ( $W/U_{ref}$ ) and lateral ( $V/U_{ref}$ ) velocities.
The dimensionless standard deviation (e.g. $U_{std}/U_{ref}$ ), kurtosis (e.g. $U_{kurt}/U_{ref}$ ) and skewness (e.g. $U_{skew}/U_{ref}$ ) of each the given (e.g. $u$ ) velocity component.
The intensity of turbulence of the given velocity (e.g. $I_{U}={\sqrt {\overline {u'^{2}}}}/U$ )
The dimensionless mean vertical ( ${\overline {u'w'}}/U_{ref}$ ) and lateral ( ${\overline {u'v'}}/U_{ref}$ ) momentum fluxes.
The dimensionless mean concentration ( $C^{*}=CU_{ref}HL/Q$ , where $C$ is the mean concentration in ppm, $H$ is the reference building height in m, $L$ is the length of the line source in m and $Q$ is the volumetric flow of the pollutant, ethane, in ml s^-1).
The standard deviations ( $C_{std}^{*}$ ), kurtosis ( $C_{kurt}^{*}$ ) and skewness ( $C_{kurt}^{*}$ ) of each the dimensionless concentration.
The dimensionless mean vertical ( $C^{*}W/U_{ref}$ ) and lateral ( $C^{*}V/U_{ref}$ ) advective pollution fluxes.
The dimensionless mean vertical ( ${\overline {c^{*'}w^{'}}}/{U_{ref}}$ ) and lateral ( ${\overline {c^{*'}v^{'}}}/{U_{ref}}$ ) turbulent pollution fluxes.
The dimensionless mean vertical ( ${\overline {c^{*}w}}/{U_{ref}}$ ) and lateral ( ${\overline {c^{*}v}}/{U_{ref}}$ ) total pollution fluxes.

Measurement techniques

When measuring the turbulent pollutant fluxes at the openings of the street canyons, the corresponding velocity component and the pollutant concentration at the respective point were measured simultaneously. This means that for the upper and lateral openings, the vertical and lateral velocity components were measured respectively. This measurement was carried out using a combination of LDA (Laser Doppler Anemometry from Dantec Dynamics A/S) and FFID (Fast-Response Flame Ionisation Detector, HFR400, from Cambustion Ltd).

To ensure proximity between the LDA measurement volume and the inlet of the FFID sampling tube, the LDA and FFID probes were mounted together on a 3D traverse system. The position of the inlet of the FFID sampling tube was carefully adjusted to be 1 mm above, 1 mm behind and 1 mm beside the centre of the LDA measuring volume (Fig 8). Through various test measurements with different probe positions, we confirmed that the influence of the FFID sampling tube on the LDA measurement was negligible.

During the measurement campaign, the LDA sampling frequency was kept between 0.5 and 1 kHz, depending on the flow range investigated. The FFID sampling frequency was set to 0.5 kHz to achieve the desired response time of 2 ms. However, due to the physical characteristics of the FFID sampling tube (length of 200 mm and diameter of 1.2 mm), an average time delay of about 12 ms was obtained in contrast to the LDA. This time delay varied depending on the air density and the dynamic pressure at the sampling point. To obtain a more accurate individual FFID time delay, we used the maximum correlation coefficient between the time series of the velocity component and the concentration. This resulted in an adjusted time delay between 10 and 13 ms.

To account for the influence of the LDA oil seed particles (with a diameter of about 1 μm produced by smoke generator Tour Hazer 2 from Smoke Factory) on the FFID concentration measurements, a correction was applied during the FFID calibration process. The measurements of the background concentration of the LDA seed particles were performed separately and subtracted from the measured calibration gas without activating the line source.

Figure 8: Snapshot of the experimental setup for the point measurement of turbulent vertical pollution fluxes in the case of the oblique wind direction.

Contributed by: Štěpán Nosek — Institute of Thermomechanics of the CAS, v. v. i.

Front Page

Introduction

Review of experimental studies

Description

Experimental Set Up

Measurement Quantities and Techniques

Data Quality and Accuracy

Measurement Data and Results