Abstract
In the structural design of tall buildings with long natural period and low damping capacity, it is very important how to estimate the dynamic effects induced by the oscillation due to wind turbulences. Theoretically wind-excited vibrations are classified into the buffeting-type forced vibration, the Aeolian vibration due to Karman voltices, the galloping-type and the flutter-type self-excited oscillation. Since realistic responses of a building may be a complex of these typical types of vibrations, purely theoretical approaches are still far insufficient to meet the realistic problem and the experimental observations are inevitably necessary. This is a report on a wind tunnel test on reduced scale models of two tall buildings. The model is situated in the center of the regional model and is exposed in a simulated turbulent wind flow. The model consists of two parts : rigid part with the exterior appearance similar to that of the prototype, and the flexible spring system with damping mechanism located at the bottom of the rigid part, which enables the rigid part to vibrate in flexural manner into two directions and torsional manner. The models were proportioned to satisfy the similarity law in aerodynamics : The time scale 1/λ_t and the length scale 1/λ_t were determined as 1/λ_t=1/34.2, 1/λ_t=1/200 for Model A, 1/λ_t=1/44.6, 1/λ_t=1/300 for Model B. Model A is 120cm in height and has almost square section (18cm×17cm) and its fundamental natural period of flexural vibration is 0.16sec., which corresponds to 5.4sec. of prototype's. Model B is 40cm high with rectangular section (24cm×15cm) and its fundamental natural period of flexural vibration in x-direction and in y-direction are 0.081sec. and 0.092sec. and that of torsional vibration is 0.097sec. which correspond to 3.30, 3.45, 4.14sec. of prototype's respectively (see Table 1). The wind tunnel flow was adjusted to satisfy the similarity to the natural wind as for the vertical profile of the mean wind speed and the turbulence intensity. Applied exponents of the power law for the mean wind speed profile were as followings. n=1/2.5 for Model A n=1/3.5 for Model B The turbulence intensity is about 8〜10% near the tip of the model and about 15〜20% at z=10cm near the ground. The configuration of the power spectra of the wind tunnel flow was almost similar to that of the natural wind, but was somewhat shifted to the higher frequency side. The results of the test are summerized as followings. 1) Among various kinds of vibrations, the most predominant mode of vibration was that of buffeting type. This fact well agrees with informations obtained by observations in the sites of existing tall buildings. Face-on winds produce greater responses of the mean, maximum, and R. M. S. displacements than the other direction winds do. It can be estimated that the maximum responses under the 60m/sec. wind reach remarkably large values; relative rotation ψ_<max>=5×(10)^<-3> for the building A and ψ_<max>=2×(10)^<-3> for the building B in the flexural mode. 2) Outlined features of the responses were well explained by the Davenport's eq・(6) which covers the buffeting type of vibration. 3) The Aeolian vibration, which did not appear in the turbulent flow, was observed evidently for Model A in the uniform flow at the range of the reduced wind velocity U=9. This range of the critical value has a good correspondence to the recent reports by Vickery. 4) The dynamic responses of the tall building are very susceptible to the differences of the mechanical properties and the shapes of the building. Several typical differences of the response of the two buildings are compared in Table 3.
