Susumu Oikawa

Wind tunnel experiments on how thermal stratification affects flow in and above urban street canyons

The effects of atmospheric stability on flow in urban street canyons were studied using a stratified wind tunnel. We conducted experiments using a model that represented city streets with simply shaped block forms, while varying atmospheric stability across seven stages from stable (Rb=0.79) to unstable (Rb=−0.21). We used a laser Doppler anemometer (LDA) and a cold wire to measure the flow field and temperature within and above the street canyon. In addition to mean values of wind speed components and temperatures, we measured turbulence intensity, shear stress, and heat flux distribution. Our results led to the following conclusions: Cavity eddies that arose in the street canyon tended to be weak when the atmosphere was stable and strong when unstable. Stable atmospheric conditions led to a positive feedback effect in which the downward flow into the street canyon weakened due to buoyancy, which facilitated the formation of a more highly stable stratification. As a result, when stability exceeded a certain threshold (somewhere in the range of Rb=0.4–0.8), the wind speed in the street canyon dropped nearly to zero.

Field and wind tunnel study of the flow and diffusion around a model cube—I. Flow measurements

Abstract Field and wind tunnel studies of the flow and diffusion around a cube were conducted. Part I of this study investigated the flow patterns behind the model. In the field study, the nondimensionalized cavity wake length (the longitudinal distance from the model center to the flow reattachment point normalized by the model height H), Lc/H, was determined using 61 small vanes arranged on the cube roof and nearby ground. Eighty three field tests were conducted and it was found that as the wind angle increased, Lc/H increased, and as the upwind turbulence intensity increased, Lc/H decreased. To isolate the most significant parameters for Lc/H, a statistical analysis was performed and it was found that wind direction had the largest positive correlation followed by Returb, while u∗/U had the largest negative correlation followed by σv/U, σu/U and σw/U. For the range of stabilities tested (−0.54 ⩽ Rif ⩽ 0.13), thermal stratification had a negligible effect on the value of Lc/H. Among these variables, the contribution of the wind direction was 62 % while that of the crosswind turbulence intensity was 12% according to a step-by-step multiple regression analysis. To clarify the results found in the field, a scale model of the cube was investigated in the wind tunnel, changing wind direction and upwind surface roughness. Similar results were noted and comparing the two studies it was found that when wind direction and upwind turbulence intensity, preferably σu/U, σv/U and u∗/U , were matched, the values of Lc/H in the field and wind tunnel agreed.

Field and wind tunnel study of the flow and diffusion around a model cube—II. Nearfield and cube surface flow and concentration patterns

Part II of a study on the flow and diffusion around a cube considered the concentration and flow patterns on and around the cube. Nonbuoyant tracer gas was released from the center of the cube roof at low exhaust velocity. The concentration patterns were shown to be strongly dependent upon the flow patterns, and in particular, on the existence of reverse flow at the source location. In the field, vane measurement of the flow showed that there was reverse flow at the rooftop center when the wind angle was small. As upwind turbulence intensity increased, the wind angle at which reverse flow occurred decreased. Modeled in the wind tunnel, this configuration was investigated for five different wind angles (θ = 0–45°) and four different upwind roughnesses (very smooth to very rough), and similar tendencies were found. Reverse flow at the source position resulted in high concentrations at the leading edge of the cube for the case of the smooth upwind surface (low upwind turbulence intensity) and θ = 0°. For the very rough surface and θ = 0°, there was no reverse flow at the source location and so the maximum rooftop concentrations occurred downwind of the source. The effect of increasing upwind turbulence was a decrease in the measured concentrations C∗. In terms of concentration similarity for the downdraft phenomenon, it was found that the wind tunnel model tended to overestimate the concentration found on the roof and may therefore be applied as a conservative indicator of the prototype. However, the same similarity criteria showed that, in addition to simulating Lc/H (see Part I), the location and value of the maximum ground level concentration was in reasonably good agreement in the field and wind tunnel.

A field study of diffusion around a model cube in a suburban area

To investigate diffusion around a building in a suburban area, a field observation was conducted on a model cube with a centrally located rooftop level source in September 1992 in Sapporo, Japan. The results show that high concentrations were observed both upwind and downwind of the source on the roof, although the mean velocity U was positive. The values of normalized concentration at locations upwind and downwind of the source were lower than those obtained from wind tunnel data conducted at moderated turbulence levels. At ground level, the mean concentrations along the model centre line show the highest value near the cube and decay rapidly in the downstream direction. The relationship between the instantaneous concentrations and instantaneous velocity was investigated using two fast-response concentration detectors and an ultrasonic anemometer. It was found that when reverse flow occurred on the roof, the tracer gas was detected upwind of the source.