EXPERIMENTAL AERODYNAMIC STUDIES ON TALL RECTANGULAR BUILDINGS

Abstract

Introduction In the last century, development in the analysis and design of structures, advancement in construction technology, and improvement in making of light weight materials with high strength has led to construction of tall buildings. With the buildings becoming progressively taller, lighter and more flexible, the vibration of tall buildings under wind became a major concern for the structural engineer. The response of tall buildings to wind depends on many parameters, such as natural wind characteristics, surrounding environment, building size, shape, orientation, and its dynamic properties (mass, damping and stiffness). The structural engineer has to ensure smaller vibration levels, particularly for the more frequent smaller wind storms and winds, while at the same time using lesser size of structural members for economical design Although the building location, its orientation and overall geometry are often pre-determined from space availability and/or architectural considerations, attempt has also to be made to reduce its response by modifying the governing parameters. Tall buildings are expensive structures. However, the extra cost involved in the prediction of their response under wind loading from wind tunnel measurements is relatively small while being very useful for arriving at a safe and economical design with acceptable levels of vibrations induced by wind. Wind tunnel investigations are needed due to shortcomings and limitations of the analytical procedures and also because very little is available on the effect of interference from adjoining tall structures. Particularly, there is no generalized approach for predicting the acrosswind response of tall buildings. In the absence of a sufficiently large number of field data on prototype measurements, wind tunnel tests provide a viable alternative. A simulated wind tunnel experiment on an appropriate model of the building is not only a powerful tool to investigate the effect of one or more parameters but also offers the advantage that it can be repeated as many times as desired under controlled laboratory conditions. Moreover, some of the aeroelastic instabilities, such as lock-in phenomenon, galloping, etc, which could be disastrous, are predictable through wind tunnel tests. Further, their elimination can also be investigated economically by wind tunnel tests. For the present work, a tall rectangular building of 300m height and 50mx25m plan size (proportions 12:2:1) was chosen for detailed wind tunnel study and generation of acrosswind response spectra, besides examining the parameters involved in the alongwind response of the building. Objectives The objectives of the present work have been as follows: 1) To study the behavior of a tall rectangular building in isolation for different flow conditions to obtain the 'basic' responses. 2) To find the effect of turbulence characteristics on the response of isolated buildings. 3) To create non-dimensional acrosswind force spectra for the building, with different conditions of the approach terrain. 4) Investigation of lock-in phenomenon in tall rectangular buildings. 5) Study of interference effect from a single tall rectangular interfering building. 6) Study of interference effect from a pair of tall rectangular interfering buildings-one upstream and the other downstream of the principal building. The Study The present study has been made in two parts- namely, analytical and experimental. In the analytical part, the Davenport Gust Factor method has been used for determining the response in the alongwind direction and Saunders' Non-dimensional Force Spectra for the acrosswind direction. This study has been carried out in isolated condition for four types of terrain conditions. The study has been made for long and short afterbody orientations. The experimental part, consisted first of generating various flow conditions; four atmospheric boundary layer (shear) flows were generated with the mean velocity variation expressed by power law exponent a as 0.12, 0.18, 0.24 and 0.30. These have been denoted as BL1, BL2, BL3, and BL4 flow respectively. For a study of the effect of turbulence characteristics on the response of isolated building, three grid generated flows were also established. These have been denoted as BG1, BG2, and BG3 flow, and their integral scales of turbulence are 6 cm, 9 cm and 18 cm. In the second stage, response of the building has been measured in both isolated and interference cases for all the seven flow conditions. Displacement and acceleration at the top of the building model were recorded in both alongwind and acrosswind directions for long as well as short afterbody orientations. Statistical parameters of the responses, i.e., Mean and RMS alongwind response and RMS acrosswind response have been computed in all the cases. In isolated condition, all the responses have been expressed as a function of reduced velocity in the form: RESPONSES x(REDUCED VELOCITY)" Where "C" and "n" depend on building shape & size and the approach flow characteristics. The alongwind and acrosswind responses have been studied as follows: 1. Analytical responses in boundary layer shear flows. 2. Comparison of analytical values with experimental values obtained in boundary layer shear flows. 3. Experimental responses in all the seven flow types as mentioned above. The study of interference effect was performed in the boundary layer flow type BL4. For this purpose, the aeroelastic model was placed in long and short afterbody orientations. Response of the model in the case of interference from a single as well as a pair of interfering buildings of same height and different plan dimensions was recorded in both the principal directions. Measurements were made for seven reduced velocities in the range 2.31 to 16.54. The three interfering buildings have proportions 1:1.5:12, 1:2.5:12, and 1:4:12 and these have been denoted by Ml, M2, and M3 respectively. The height of all the three interfering buildings was kept the same as that of the principal building. The response of the model in the IV case of interference from a single building has been presented as buffeting factor contours covering all interfering locations. The buffeting factor (BF) has been defined as: BF= Response ofthe model in interference case / Response ofthe model in isolated case The following interference cases have been studied in the present work: i) Interference due to presence of a single interfering building in upstream or downstream location a) Both principal and interfering buildings in long afterbody orientation b) Both principal and interfering buildings in short afterbody orientation c) Principal building in short afterbody orientation and interfering building in long afterbody orientation ii) Interference due to presence of apair of interfering buildings one in upstream and the other in downstream position a) Both principal and interfering buildings in long afterbody orientation b) Principal building in short afterbody orientation and both interfering buildings in long afterbody orientation Conclusions The outcome of the presentwork is summarized as follows: 1) Relationships between building response (in both the principal directions) and reduced velocity have been established for all flow conditions for long as well as short afterbody orientations. 2) Agood agreement has been found between analytical and experimental results for the isolated building. 3) Based on the results of the present study a relationship between terrain drag coefficient and power law exponent has been established. 4) The RMS acrosswind response is found to increase more than the RMS alongwind response with increase in the turbulence parameter Iu(Lux/b)2, where Iu is turbulence intensity, Lux is integral scale ofturbulence and bis width of the building. 5) Non-dimensional acrosswind force spectra has been generated in simulated atmospheric boundary layers for four types of approach terrains for a tall building of 1:2:12 proportions and 300m high. 6) The non-dimensional acrosswind force spectra thus generated can be used to predict the acrosswind acceleration and corresponding overturning moments at the preliminary design stage oftall rectangular and square buildings as well as at the final design stage of similar buildings. 7) This study has revealed that the acrosswind response oftall rectangular building studied dominate over its alongwind response. 8) Lock-in phenomenon has been observed to occur for the short afterbody orientation in all flow conditions. Interestingly, this phenomenon gets eliminated when the building is placed in the vicinity oftwo tall interfering rectangular buildings. 9) A comparison between analytical and experimental RMS acrosswind response was made and it has been found that the non-dimensional acrosswind force spectra available in literature could be usefully employed for the preliminary design of tall rectangular buildings. The peakresponse varied between 3.5 and 4.5 of the RMS response, except for the 'Lock-in' situation when resonant condition developed so that the vibrations had a single harmonic and the peak was 1.5-2.0 times the RMS value. 10) In the case of interference from a single interfering building, the RMS alongwind response of the principal building decreased with increase in the size of the interfering building for all three combinations of the relative orientation of the two buildings (principal and interfering). 11) When the principal building was placed in short afterbody orientation, the RMS acrosswind response due to interference from a single interfering building increased with increase in the size of the interfering building for either orientations of the interfering building (that is, long and short afterbodies). The reverse was found to happen when both principal and interfering buildingswere in the long afterbody orientation. 12) In the case of interference from a single interfering building, the maximum BF value for Mean alongwind response occurred for the interfering building model M2 having dimensions close to the principal building. 13) In the case of interference from a pair of interfering buildings, all responses decreased with increase in the size of the interfering buildings. 14) Generally, the buffeting factor increased with increase in the wind speed and decreased again after reaching a peak value. 15) The behaviour of principal building in short afterbody orientation was found to be much different from that in the long afterbody orientation, both in isolated as well as in the interference case. VI