Abstracts:Emergency response forces need to be prepared for making vitally important decisions in the possession of very few trustable information. Chemical plants using hazardous materials are usually monitored for accidental release, equipped with local wind sensor and some estimates are available for the potential source intensity. Concentration levels are expected to be forecasted from this minimum information in case of accident. Most fast response dispersion models rely on the Gaussian dispersion model, despite it being limited to small scale turbulence induced by surface shear. In this paper, data of multiple field experiments are brought to common platform and, based on the statistical investigation of this dataset, an atmospheric dispersion model is proposed, which allows for the reconstruction of the pointwise concentration statistics as functions of windward and lateral coordinates. The new statistical model may be used to predict the ground level concertation profile of given probability, with the limitations to near ground continuous sources localized on flat terrain at medium latitude. At least an order of magnitude higher lateral spread of the plume can be seen compared to standard Gaussian plumes and the profiles are tilted to the right reflecting the effects of large scale atmospheric turbulence and Coriolis effects.

Abstracts:In the present study, turbulent airflows and particle deposition around two inline buildings were investigated numerically. A computational modeling approach including the RNG k-ε and Realizable k-ε turbulence models were used for these simulations. The Lagrangian particle tracking approach was implemented for evaluating dispersion and deposition of spherical dust particles. For simulation of turbulence fluctuations, an improved discrete random walk (DRW) model that includes the near wall correction was implemented into several user-defined functions (UDFs) that were linked to the ANSYS-Fluent code. It was shown that the new improved DRW stochastic model led to results that are more accurate compared to the standard model. The improved DRW model was then used for simulating turbulent deposition and dispersion of dust around building models. The presented results showed that putting buildings on elevated supports reduces dusts deposition from downstream of single and inline buildings at short distances, particularly for small particles of about 1  μm. It is also shown that particle deposition fractions around buildings on rough ground are higher than those for smooth ground. Deposition fractions were also predicted by the improved model for faces of single and inline buildings.

Abstracts:The selection of turbulence models has always been one of the important aspects of computational fluid dynamics model simulations of wind flow and pollutant dispersion problems. The commonly industrially adopted steady Reynolds-averaged Navier–Stokes turbulence models are the realizable $k-\epsilon$ model and Menter's shear stress transport $k-\omega$ model.

Abstracts:Correct prediction of the recovery of wind turbine wakes in terms of the wind velocity and turbulence downstream of the turbine is of paramount importance for the accurate simulations of turbine interactions, overall wind farm energy output and the impact to the facilities downstream of the wind farm. Conventional turbulence models often result in an unrealistic recovery of the wind velocity and turbulence downstream of the turbine. In this paper, a modified k – ω turbulence model has been proposed together with conditions for achieving a zero streamwise gradient for all the fluid flow variables in neutral atmospheric flows. The new model has been implemented in the simulation of the wakes of two different wind turbines and the commonly used actuator disk model has been employed to represent the turbine rotors. The model has been tested for different wind speeds and turbulence levels. The comparison of the computational results shows good agreement with the available experimental data, in both near and far wake regions for all the modeled wind turbines. A zero streamwise gradient has been maintained in the far wake region in terms of both wind speed and turbulence quantities.

Abstracts:The effects of upstream turbulence on roof pressure fluctuations of a low-rise building are investigated via the quasi-steady (QS) vector model. Atmospheric boundary layer (ABL) turbulence, with intensities ranging from 13% to 27% and integral length scales from 6 to 13 times the building height, is simulated in a boundary layer wind tunnel. The model building surface pressures are measured synchronously with the velocity at a point one building height above the leading edge. The QS model is found to accurately explain the effects of the ABL turbulence with scales larger than about 5 building heights. Furthermore, a QS model established in one terrain can explain the pressure fluctuations in the other terrains based on the fact that the model functions are similar over the range of observed upstream terrain conditions. This finding is important to the Partial Turbulence Simulation approach since there may be a range of turbulence intensity – integral scale combinations which can yield the same aerodynamic behavior, allowing more flexibility for choosing the model length scales. However, further work is required for modelling the effects of small-scale turbulence on the peak pressures and in defining appropriate bounds when the precise turbulence simulation can be relaxed.

Abstracts:A cycling peloton is the main group of cyclists riding closely together to reduce aerodynamic drag and energy expenditure. Previous studies on small groups of in-line drafting cyclists showed reductions down to 70 to 50% the drag of an isolated rider at same speed and these values have also been used for pelotons. However, inside a tightly packed peloton with multiple rows of riders providing shelter, larger drag reductions can be expected. This paper systematically investigates the drag reductions in two pelotons of 121 cyclists. High-resolution CFD simulations are performed with the RANS equations and the Transition SST-k-ω model. The cyclist wall-adjacent cell size is 20 μm and the total cell count per peloton is nearly 3 billion. The simulations are validated by four wind-tunnel tests, including one with a peloton of 121 models. The results show that the drag of all cyclists in the peloton decreases compared to that of an isolated rider. In the mid rear of the peloton it reduces down to 5%–10% that of an isolated rider. This corresponds to an “equivalent cycling speed” that is 4.5 to 3.2 times less than the peloton speed. These results can be used to improve cycling strategies.

Abstracts:In recent years, governmental legislation and consumer demands are driving the development of more energy efficient road vehicles. One of the aspects considered when increasing efficiency is the aerodynamic performance of the vehicle. The focus of this paper is on wake effects for a vehicle with tapered rear extensions in side wind conditions. For this purpose, numerical simulations are analysed using post-processing techniques such as Proper Orthogonal Decomposition (POD) and Two-point correlation. The extensions protrude 150 mm from the perimeter of the base and are investigated in two configurations: with a smooth taper and with an added kick. The kick realigns the perimeter base flow to the vehicle's driving direction.

Abstracts:Downwind two-bladed rotor configurations can have advantages in reducing rotor mass for wind turbines, compared with three-bladed upwind designs. However, the tower shadow adds an aerodynamic complication that can be difficult to quantify and predict. This study presents and analyzes a previously unpublished subset of data collected during an extensive wind tunnel campaign, the Unsteady Aerodynamic Experiment (UAE). At high tip speed ratios, the tower shadow is a dominating contributor to bending moment oscillations but can be mitigated by the use of a tower fairing when such a fairing is aligned with the flow. At lower tip speed ratios where the blades can undergo aerodynamic stall and hysteresis, tower shadow was only a secondary contributor to bending moments and the tower fairing did not significantly impact bending moments. The aeroelastic simulation code called FAST was used to predict the same experimental conditions. In general, simulations reasonably predicted most of the cycle-averaged aspects, but only qualitatively predicted the unsteady variations due to tower shadow. To improve simulation predictions inside the tower wake, it is suggested that future work model the unsteady wake component associated with cylinder shedding and to consider a wake model for tower fairings at various wind incidence angles.

Abstracts:High-resolution Weather Research and Forecasting (WRF) simulations of surface wind, pressure, temperature, humidity and snow depth in Arctic coastal regions are performed in this study. We compare simulation results for different elevation resolutions, surface roughness lengths, PBL schemes, initialization data and snow cover conditions with meteorological observations and validate WRF's usage in high latitudes. WRF performs well in simulations of high-latitude strong wind if the roughness length is carefully modified considering site's real vegetation type, though WRF produces insufficient wind fluctuations. Noah LSM is effective in updating surface roughness length according to snow depth. Higher elevation resolution increases surface pressure accuracy. MYNN3/YSU PBL scheme is recommended for wind speed/temperature simulation. In addition, the vertical structure of the Arctic atmospheric boundary layer is discussed, and a statically stable and turbulent surface layer is the Arctic winter is reported.

Abstracts:The long-span bridges, frequently located in the coastline and mountainous areas, are prone to suffering from the transient downburst winds associated with the thunderstorms. Therefore, it is important to examine the behaviors of long-span bridges under such wind events. In this study, the time history of non-turbulent downburst wind field is modeled using the impinging jet-based computational fluid dynamics (CFD) approach. It has demonstrated high-fidelity simulations of the downburst wind field, especially for the near-surface radial winds that are particularly significant for present investigation of bridge aerodynamics and aeroelasticity, with a reasonable computational cost. The simulated time-varying mean wind velocity field of a typical downburst event is validated using available field measurements. The correlated nonstationary fluctuations of the downburst are stochastically simulated by the Hilbert-wavelet scheme based on full-scale data, and then superimposed onto the CFD-based transient mean wind field. The obtained turbulent downburst wind field is employed as the dynamic inputs to the line-element-based three-dimensional (3-D) finite element model of the long-span bridge, and its downburst-induced transient response is acquired using the computational structural dynamics (CSD) approach. The aerodynamic and aeroelastic couplings between the downburst winds and the long-span bridge are modeled using the two-dimensional (2-D) indicial response functions that could well represent the transient bridge aerodynamics under non-synoptic winds. The time-domain buffeting response analysis of a long-span suspension bridge under the traveling downburst has been carried out utilizing the CFD-CSD-based hybrid methodology. The results highlight the importance of the transient and nonstationary effects on the non-synoptic wind-induced structural dynamic response.