Wind loads play an important role in the design and construction of long-span bridges, especially for that built in mountain valley, and wind characteristics in mountain terrain are still unclear. To further understand the wind characteristics at a bridge site in a mountain valley, large eddy simulations were employed to comprehensively study the wind characteristics at the bridge site in an L-shaped canyon. Different wind directions consisting of 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315° were considered, and the wind characteristics on the main girder and towers, as well as the flow field, were comprehensively investigated. The results indicate that the mean and fluctuating wind characteristics are significantly different from those specified in the relevant specifications, and it is inappropriate to apply the specifications directly to mountainous terrain. When wind flows along a channel, it leads to a large wind speed amplification factor, strong non-Gaussianity, and large kurtosis and skewness. The steeper the slope of the mountain, the stronger the non-Gaussianity of the fluctuating wind. The coherence function remains nearly unchanged when the distance between the two points is constant, regardless of the location. The wind characteristics at the main girder and towers are so different that they should be considered separately and taken into account in engineering practice. This study provides a reference for the wind-resistant design of bridges and structures built on mountainous terrains.

Buffeting is a kind of wind-induced vibration phenomenon of long-span cable-stayed bridges which is easy to occur. Balanced cantilever method is widely employed for the construction of cable-stayed bridge. The maximum double-cantilever state is the most dangerous state of wind resistance. Through the aeroelastic model wind tunnel experiment, the buffeting response of a cable-stayed bridge in construction under skew winds is studied in this paper. The variation law of buffeting response with the wind yaw angle is compared between the two cases with and without pylon. The experimental results show that for the same wind yaw angle the lateral and vertical buffeting responses of the main girder increase with the wind speed approximately in a quadratic curve, while the torsional buffeting response tends to increase linearly. At the same wind speed, the buffeting response of the main girder presents non-monotonic changes with increased wind yaw angle. Under skew winds, due to the disturbing effect of the pylon and the influence of the three-dimensional flow field, the buffeting response of the main girder will be affected by the aerodynamic interference of pylon. When the wind yaw angle is small, the interference effect is not obvious, with the increase of wind yaw angle and wind speed, the interference effect begins to appear. Generally speaking, the aerodynamic interference of the bridge pylon on the buffeting responses can be ignored.

This study investigated the statistical properties of the pressure fluctuations on a square cylinder across three distinct turbulence fields characterized by varying turbulent integral scales. The effect of turbulent integral scale on the non-Gaussian characteristics and extreme surface wind pressure acting on square cylinders beneath the separating flow were studied in detail. The findings indicated that the pressure distribution on the windward surface generally conformed to a Gaussian distribution, whereas notable non-Gaussian characteristics were observed in the pressure distribution on the side and leeward surfaces. The fluctuating pressure, skewness, kurtosis, peak factor, and extreme pressure increase with an increasing ratio of turbulent integral scale to structural depth (Lux/D), whereas the mean pressure remains unaffected by variations in Lux/D. As Lux increased, the energy of the internal vortices in the shear layer also increased. As a result, the non-Gaussian features of the pressure caused by vortex breakdown become more pronounced. Compared with Lux/D = 1.96, the underestimated value of the extreme pressure on the square cylinder had a maximum difference of up to 15.4% at Lux/D = 0.53. Therefore, the corresponding turbulent integral scale should be accurately simulated when measuring wind loading on a structure through wind tunnel tests.