Small disturbance potential theory is widely used in solving aerodynamic problems with low Mach numbers, and it plays an important role in engineering design. Concerning structure wind engineering, the body of the structure is in a low velocity wind field, with a low viscosity of air and thin boundary layer, therefore, the tiny shear stress caused by the boundary layer can be ignored, only wind pressure being considered. In this paper, based on small disturbance potential theory, the fluid-structure interaction between the wind and membrane structure is analyzed by joint utilization of the boundary element method (BEM) and finite element method (FEM) through a loose-coupling procedure. However, the boundary of flow field to be calculated is not fully smooth, corners and edges still exist, so the discontinuous boundary element is introduced. Furthermore, because a large scale boundary element equation set with a nonsymmetrical coefficient matrix must be solved, this paper imports a preconditioning GMRES (the generalized minimum residual) iterative algorithm, which takes full advantage of the boundary element method. Several calculation examples have verified the correctness and soundness of the treatments mentioned above.
LI Ya &YE JiHong Key Laboratory of R.C. & P.C. Structures of Ministry of Education, Southeast University, Nanjing 210096, China
The wind pressure characteristics on a saddle roof at wind direction along the connection of the low points are systematically studied by the wind tunnel test. First, the distributions of the mean and the fluctuating pressures on the saddle roof are provided. Through the wind pressure spectra, the process of generation, growth and break down of the vortex on the leading edge is presented from a microscopic aspect and then the distribution mechanism of the mean and fluctuating pressures along the vulnerable leading edge is explained. By analysis of the wind pressure spectra near the high points, it can be inferred that the body induced turbulence reflects itself as a high-frequency pressure fluctuation. Secondly, the third-and fourth-order statistical moments of the wind pressure are employed to identify the non-Gaussian nature of the pressure time history and to construct an easy tool to localize regions with a non-Gaussian feature. The cause of the non-Gaussian feature is discussed by virtue of the wind pressure spectra. It is concluded that the non-Gaussian feature of the wind pressure originates from the effects of flow separation and body-induced turbulence, and the former effect plays an obvious role.
The concept of the coherence function is adopted to find the wind pressure correlation of two points on domes of different rise-span ratios. The pressure measurements are made on the dome roof models by the wind tunnel test. The coherence functions for different separation distances at several directions of the domes from different wind directions are examined. The results show that there is a strong correlation for two adjacent points at low frequency, but not for non-adjacent points. The coherence of the wind pressure increases with the decrease in the separation distance. Moreover, the coherence of the wind pressure is in the strongest correlation on the along-wind direction at the same separation, but the lowest correlation is on the cross-wind direction. The detailed derivation of the proposed exponential coherence model of the wind pressure from experimental data is also discussed. It is found that the proposed exponential coherence model can be appropriate, especially, for small separations and the change in the directions on domes. Based on the quasi-steady theory, the relationship between the wind pressure and the wind velocity on the basis of the coherence model is also examined. The coherence observed between the wind pressure and the wind velocity is not adequately predicted by the quasi-steady theory.
