Turbulence and channeling in a simple urban environment

Urban boundary layer simulations were performed in the water channel at the University of California, Riverside, in the Laboratory for Environmental Flow Modeling. A recirculating water channel was utilized to create controlled flow field. The channel can produce a maximum mean velocity of 0.5 m/s in the test section. A variable frequency controller allows pump control with a resolution of 1/100 Hz which corresponds to the mean velocity change of 0.1 mm/s. Building configurations simulating simple urban patterns were accomplished using highly polished acrylic models to minimize effects of refraction and attenuation of the laser sheet utilized by the TSI Particle Image Velocimetry (PIV) system. Using a simple two building configuration the effects of channeling were studied and distribution of the turbulent kinetic energy was measured. Flow velocity was varied from 2 cm/s to 10 cm/s, corresponding to Reynolds Numbers from 2,000 to 10,000. Flow approach angle was 1, 3, 5 and 7 degrees and the investigated ratios of building heights to the street width were 0.5, 2 and 4 corresponding to the skimming flow, wake interference, and isolated wake regimes, respectively. By utilizing TSI's PIV system, resultant flow fields were measured over a period of up to 5 minutes. The occurrence of flow channeling vs. flow recirculation was observed throughout experimentation. The criteria for channeling occurrence in this 2 building configuration was established. When the flow approach angle is greater than the critical value of ~5 degrees, for building setups in the skimming regime, it was noticed that the flow velocity no longer affects the channeling occurrence and channeling is always observed. In future experiments, critical approach angle will be precisely defined by refining the angle step size for each of the three regimes. For the flow approach angles smaller than the critical angle, increase of flow velocity will eventually lead to the channeling in all cases. Here the channeling occurrence is a function of both the flow angle and flow velocity. As a third varied parameter, the influence of building spacing was investigated. As expected, the wider building spacing was more prone to the channeling. In a phase that exists between pure channeling and no channeling, the recirculation regions and vortex shedding is observed. The distribution of turbulent kinetic energy (TKE) around the buildings was related to the building geometries and approaching flow conditions. These new parameterizations of TKE are to be implemented into the in-house developed urban dispersion model. The ongoing experiments will further cover more complex geometries and the influence of several buildings.