WIND TUNNEL STUDIES ON AERODYNAMIC INTERFERENCE IN TALL RECTANGULAR BUILDINGS

Abstract

Wind induced response of a tall rectangular building is a function of many parameters. These include the geometric and dynamic characteristics of the building as well as the turbulence characteristics of the approach How. The flow approaching a tall building inherits the combined effect of the wake of obstacles it has overcome. These may be natural features like forests, hills etc., and also man-made structures both low and high rise buildings. A few analytical approaches are available for the estimation of the wind induced response of tall buildings in along and acrosswind directions. These estimates can only be verified by prototype measurements and in the absence of a sufficiently large number of field data, wind tunnel tests provide a viable alternative. A simulated wind tunnel experiment on an appropriate model of the building yields results which give a deeper insight into the phenomenon and provides more precise information, overcoming the shortcomings of the analytical formulation. Besides studying interference effects, an effort has also been made in the present work to study some of the factors governing the response of isolated tall rectangular buildings. A building of width:depth:height proportions as 1:2:10 has been chosen for the study. In the first phase of the study wind tunnel experiments have been performed in boundary layer flows and also in grid generated flows on a 'stick' type aeroelastic model which simulates the fundamental mode shape of vibration of the building in either direction. Displacements and accelerations at top of the model were observed for a reduced velocity range of 3 to 18. The building being rectangular, it has been studied along both of its principal directions, i.e., in long and short afterbody orientations. When a building is located in an urban environment, it is exposed to a wind of altogether different characteristics than wind over an open region. This is on account of the turbulent 'wake region' created behind an upstream building. The interaction with the upstream building(s) can produce significant changes in the responses of a tall building. While several wind tunnel studies have been carried out on square buildings to find interference effects, there are only a few studies reported on the interference effects on a tall building of rectangular section. In the second phase of this study, effect of interference on the same building due to other similar buildings, but of varying aspect ratio in plan (1:1, 2:1, 3:1), have been studied. For this purpose the aeroelastic model of the building has been used while the interfering models are rigid. The interfering models have been referred to as 'narrow', 'medium' and 'large' width models. To cover a wide range, the interfering model was placed at various locations on the upstream as well as downstream side of the aeroelastic model. Therefore, the second part of this research work has been addressed to study the effects of interference. The interference study is carried out for long and short afterbody orientations of the building under investigation for three different wind speeds. Further, some of the earlier studies have shown that the interference effects are different when a pair of buildings rather than a single one is placed upstream. Therefore, the last part of this research work has been addressed to study the effects of interference arising out of a pair of buildings on the upstream side through wind tunnel experiments. Three different interfering pairs have been considered in this part of the study. Two pairs consist of 'narrow' width rigid models and the third has 'medium' width models. The models are placed either at + 3b and -3b lateral offsets or one of the model at +3b offset with the other located centrally(in-line), to form an interfering pair. Results of the isolated building study are presented as its response related to the reduced velocity. The experimental results have been projected to estimate the full scale values using appropriate scaling laws. These have also been compared with the prototype responses computed analytically. The analytical values in the alongwind direction have been obtained using the Davenport's 'Gust Factor Approach'(1967) for slender line-like structures. In the acrosswind direction, while no general analytical formulation is available, the force-spectra developed by Saunders(1974) have been made use of, though this has its limitations, since the spectra have been developed on models with significant differences in geometric, structural and flow characteristics compared the one in this study. Results of interference study are presented in the form of 'Buffeting Factors'(BF) defined as r>p _ Response in Interference configuration Response in stand-alone configuration For the 'single interfering building' study, contours of BFs are plotted to cover the entire region around the building. The BF contours have been prepared for each of the three interfering models, and, for the various wind speeds. These contours show graphically the variation of BF for a particular response parameter for a specific 111 reduced velocity as the interfering model occupies different positions. As before, both long and short afterbody orientations have been considered. For interference by 'a pair of upstream buildings', BFs are plotted against the longitudinal distance between the pair and the aeroelastic model, for its long and short afterbody orientations, for three different wind speeds. This extensive study has yielded significant results for the isolated building as well as for the interference conditions. These are summarised as follows: (i) It is possible to express both along and acrosswind response of an isolated building as a function of the reduced velocity as follows: Response = C*(Reduced Velocity)" where constants 'C & 'n' depend upon the characteristics of the building and the approach flow. The values of 'C and 'n' have been obtained and presented for the 1:2:10 proportion building investigated in this study, adding to the data base for design of similar buildings. (ii) Acrosswind response is seen to dominate over the alongwind response. (iii) Turbulence parameters of the approach flow, i.e., 'intensity' and 'scale', have significant effect on the response. The response in alongwind direction increases with the term Iu(b/Lux) . However, the acrosswind response exhibits a peak for a value of this term lying between 4 and 5. (iv) Davenport's gust factor approach may overestimate the alongwind response unless the value of drag coefficients, correlation of surface pressures across the face of the building, and the parameter 'L' defining the gust energy ratio are carefully selected. Satisfactory predictions can be made by adopting appropriate values of these parameters in the formulation. (v) For the estimation of response in the acrosswind direction, use of available force spectra may yield good results only for buildings with similar characteristics. However, separate force spectra should be developed for buildings which differ significantly in geometric and structural characteristics. (vi) In case of interference from a single building of equal height, significant interference effects are observed with 'narrow' and 'medium' width interfering iv buildings. For rms response, BFs as high as 9 are observed in the alongwind direction and may go upto 16 in the acrosswind direction. Maximum response, however, gets enhanced 2 to 3 times. Critical zone of interference (BF > 2) extends from 4b to 10b upstream and 2b to 4b downstream, with a lateral offset upto 5b both on upstream and downstream side. Interference effects are larger in short afterbody orientation than in the long afterbody orientation. (vii) Interference effects are more pronounced with a pair of buildings than a single building on the upstream. Largest interference effects are observed for a pair of buildings when the interfering buildings are at 3b lateral offsets. Also, in this case, the alongwind response is much more enhanced than the acrosswind response. The critical interfering locations are found to lie between 4b and 10b.