Natural ventilation in buildings is driven by the pressure field across the building envelope due to variations in wind and indoor-outdoor temperature differences. For a building in an urban canopy, prediction of the air exchange rate is made difficult by the spatially varying, highly turbulent flow. This study presents an idealised, full-scale experiment that tests two effects of urban canopy flow on building ventilation rates: 1) sheltering from neighbouring buildings causing variability in the wind direction, and 2) turbulent contribution to ventilation rate. The experiment also provides full-scale data to test the effectiveness of both steady-state engineering calculations, and CFD modelling of urban ventilation as part of the REFRESH project.

To represent a simplified urban area, a staggered array of 8 cubes of height 6 m was built around an instrumented metal cube at a relatively flat, rural site in Silsoe, Bedfordshire, U.K. The experiment took place from September 2014 to January 2015, with the array removed part-way through to allow the instrumented cube to be studied in isolation. Sonic anemometers were positioned immediately upstream and downstream of the cube openings at 3.5m, and at 6m and 10m on a reference mast upwind (with respect to the prevailing wind direction) to monitor the wind-flow patterns around the array. Within the instrumented metal cube were 2 sonic anemometers, 24 thermocouples, 30 external and 2 internal pressure taps alongside up to 4 CO2 sensors. CO2 tracer gas decay experiments were used to gain an understanding of the ventilation and infiltration rates of both the isolated and sheltered cube set-ups for three ventilation configurations: sealed cube, single-sided and cross-ventilated, the latter two involving one or two square openings respectively in the centres of the front and back faces of the cube. For the tracer gas experiments, attention was paid to sensor positioning and the effect of mixing within the cube for each configuration. Ventilation rates will be presented and related to the measured flows and pressure patterns across the cube envelope.

Wind tunnels are frequently used to model urban street canyon turbulence and ventilation dynamics [1,2]. However, there have been few studies in which the mean and unsteady flow dynamics from a wind tunnel and field study have been quantified and compared in order to justify the validity of the wind tunnel results. The present work examines the flow field in a simple street canyon that has been modeled at full-scale and at 1:200 scale in a wind tunnel. Field data were provided from the Influence des effets micro-météorologiques sur la propagation acoustique en milieu urbain (EM2PAU) campaign [3], which took place in Nantes, France, over a two-year period, in which shipping crates were used to model an idealized canyon 24m long, 3.64m wide and 5.20m high. Sonic anemometers were used to measure the three components of velocity at 6 locations in a cross-section of the canyon. The wind tunnel experiment was conducted in the low-speed, suck-down boundary layer wind tunnel in the Laboratoire de recherche en Hydrodynamique, Énergétique et Environnement Atmosphérique (LHEEA) at École Centrale de Nantes. The experiments used five 800 mm high vertical tapered spires, a 200 mm high solid fence across the working section and an upstream roughness consisting of 50 mm cubes arranged in a staggered array with λp = 25% to initiate the boundary layer development. Flow measurements were conducted using stereoscopic PIV to record all three components of velocity within the canyon. Comparison between the field and wind tunnel study has demonstrated good agreement for the mean velocity and turbulence statistics including two-point correlation of the three velocity components, which are typically within 20%. However, significant differences in the along-canyon mean and turbulent components have been observed and are shown to be a result of the changing of the ambient wind direction and low frequency motion present in the field. It is also shown how the application of a Stochastic Estimation (SE) method [4], using spatially well-resolved wind tunnel Particle Image Velocimetry (PIV) measurements may be used to predict full-scale flow dynamics from the limited number of field measurement anemometers. The main statistical features of the flow, such as the presence of the shear layer that develops over the canyon and the intermittent ejection and penetration events across the canyon opening, were well predicted by the SE. This combination of SE and wind tunnel PIV measurements may be used to optimize the number and location of wind velocity measurement locations in future street canyon field measurement campaigns, but also to spatially extrapolate instantaneous flow features measured by the limited number of sonic anemometers. The full paper will discuss in detail the comparison between the field and wind tunnel mean turbulence statistics as well as the reasons for discrepancies between the two and the SE results.

REFERENCES

[1] P. Kastner-Klein and M.W. Rotach (2004) Mean flow and turbulence characteristics in an urban roughness sublayer. Bound Layer Meteorol 111:55-84.

[2] E. Savory, L. Perret and C. Rivet (2013) Modelling considerations for examining the mean and unsteady flow in a simple urban-type street canyon, Meteorol Atmos Phys 121:1-16.

[3] G. Guillaume, C. Ayrault, M. Berengier, I. Calmet, V. Gary, D. Gaudin, B. Gauvreau, Ph. L’hermite, B. Lihoreau, L. Perret, J. Picaut, T. Piquet, J.M. Rosant and J.F. Sini (2012) Micrometeorological effects on urban sound propagation: a numerical and experimental study, All Glo Sus 19:109-119.

[4] N. E. Myrray and L. S. Ukeiley (2007) Modified quadratic stochastic estimation of resonating subsonic cavity flow. Journal of Turbulence, Volume 8, Art. No. N 53.

On the countermeasures of urban heat island, improvement of the material covering urban surface and improvement of these physical shapes are studied independently. The application as an appropriate combination of these two in consideration to location (climate) and weather condition (season and time) etc. is one of key issue. About the radiation and the energy balance on the urban surface, no similarity theories like the one for the flow field exist and its control is also difficult. Therefore, there are few experimental cases using a scale model like Spronken-Smith and Oke (1999) (on urban green park) etc. Solar radiation heats the surface of the urban canopy and it is resulted that characteristic air flow pattern is formed by the strong buoyancy inside the street canyon. In a general wind tunnel experiment, scale model surface is heated by electric wires and this approach doesn't express the radiation balance on the scale model surface explicitly. In this study, the authors tried to minimize the influence of roughness to flow field and to settle halogen lamps in close position. This is a challenging approach to reproduce the similar condition of radiation balance to the summer sunny day's in the wind tunnel. When the wind velocity is small, the effect of the surface material is obviously shown in the thermal-wind environment at the vicinity of the roof top part (increase of the wind velocity and decrease of the turbulent energy).

It is well known that the airflow around a group of buildings is an important factor for urban design in considering amenity of urban inhabitants, such as thermal comfort or air quality because wind is one of the most important driving force of heat and scalar transfer in urban area. From a view point of fluid dynamics, airflow in urban area is assumed to be turbulent flow field over roughness and the physical property of it has been examined in detail.

For the last decades, numerous researchers have investigated the relation between turbulent flow field around buildings and geometrical parameters of urban models like plan area ratio (λp, a ratio of plan area of a building to unit district area). As the result of that, various useful knowledge of qualitative and quantitative property of turbulent flow field in urban area was provided.

Most of these past studies used urban models consist of highly simplified building models, like cubes or rectangular parallelepiped; however, in general, actual buildings have secondary roughness like roofs, balconies and penthouses. As a result, urban geometry become complex and it is uncertain whether the knowledge presented in the past studies is directly applicable to estimate the airflow nature of real urban setting. Actually, a few pioneering studies point out the possibility that turbulent flow field structure change when secondary roughness attached on a main building body. For example, the Computational Fluid Dynamics (CFD) simulation using Reynolds Average Navier Stokes (RANS) model (Mohamad et al. 2014) revealed the drastic effect of porch on temporal-averaged flow field around 2D building models. Furthermore, they indicated that the wind-induced natural ventilation rate inside a building significantly varies with flow field around buildings.

Under these circumstances, in this paper, the authors perform wind tunnel experiment to investigate how turbulent flow field in 2D street canyon is changed by flat eaves overhang the street. Split Film Prove (SFP) and Particle Image Velocimetry (PIV) are employed to capture spatial distributions of turbulent statistics as well as unsteady flow motions around the canopy. The two types of model configuration are adopted for the measurement; they are regularly-spaced uniform-shape square bars with and without eaves arranged face to face perpendicularly against the mean wind direction. The street aspect ratio (a ratio of height of a bar to width of street) is 0.33; thus the wake interference flow regime is expected to observe in street canyon without eaves. The measurement result will depict how the eaves modify the mean flow filed as well as the turbulent mixing within and above the 2D urban canopy.