|The thesis presents a systematic experimental and numerical study on the interactions among porous fence, airflow, and windblown snowdrifts, a knowledge that will contribute to optimize the performance of porous wind shielding system in Cold Regions. A comprehensive review of the concepts, theories, techniques, and key findings associated with the research work has been undertaken. The key technical parameters influencing fence performance have been systematically studied by means of wind tunnel experimental investigations and Computational Fluid Dynamics (CFD) simulations.
The study has found that porosity is the most influential structural parameter affecting the performance of porous fences in many aspects. Fence height stands a significant positive position in terms of its performance. It was found that fence performance is not sensitive to the changes of approaching atmospheric airflow velocity. Nevertheless, a bottom gap can improve snow fence trap efficiency. All of those findings agree with most of the findings of other researchers, which affirms that the research methodology adopted in this research is sound.
Physical experimental work was performed to assess the reliability and credibility of the numerical models. Those models have been intentionally simplified, which made them easier to construct and quicker to obtain numerical solutions at a lower computational cost. Furthermore, the numerical models demonstrate the level of competence acquired through this research that is implemented in the optimisation of fence design.
Special attention has been paid to the issues where elaborate research work has not been systematically reached in the open literature, this includes areas such as the effects of arrangement of porous holes, fence surface shear, and directions of wind load with respect to the fence, etc. Correlation between the reattachment length, the shelter distance, and the creation and distribution of fence surface shear is reported, to the author’s knowledge, for the first time in the open literature. General guidelines for the design of shelters based on porous fences have been established through this study. For example, the desirable size of hole range should be identified beforehand, and porous holes with sharp angular corners should usually be avoided in the fence design. It is recommended to place the fence within an angle of 30° to the wind load, where the effective shelter distance can be estimated in a linearized equation, and the normal drag coefficient can be described as a function of cos2θ. Optimal design of the arrangement of porous holes will maximize the fence performance, especially when the close fence environment is of concern.
Although the definition of fence effective zone is still vague in the research field, the key factors influencing the fence effective zone have been investigated by evaluating the reduction of wind velocity leeward of the fence in this thesis. It is found that the fence effective zone is not sensitive to the change of approaching airflow velocity, and that increasing fence height will increase the physical size of the fence effective zone, but not in a proportional manner. It is also concluded that fence effective zone will be significantly reduced when the non-normal wind load is inclined at an angle greater than 30° to the fence. The effective zone increases effectively when the fence porosity is optimal.
In contrast to the majority of published research work, the transient snow transport model presented in this work considers the snow transport rate as a whole without distinguishing the rate in saltation and suspension layer. The numerical study indicated that the position of the snow crest is mainly determined by the fence height, while porosity and bottom gap mainly affect the downwind deposition length. The optimal porosity for snow fences is in the range of 0.4 to 0.5, which is greater than the one for wind fences, which lies in the range from 0.25 to 0.35. Two snow crests have been observed leeward the fence at the onset of snow deposition, when the fence was placed without a bottom gap to the snow ground. This finding has not been encountered in any of the reported research work.
Wind tunnel simulations of snowdrift around the fences have marginally under-predicted the sizes of snow deposition. The numerical predictions were quantitatively and qualitatively in good agreement with the field observations. This incompetence of wind tunnel experiments on porous fences implies that numerical modelling can play a more important role in snow fence research.