Research On Wind Induced Vibration Of Long-span Roofs Using Wind Tunnel Test And Field Measurements

In recent years,more and more long-span structures have been built with increasing span and structural refinement.Roofs of such structures usually have the characteristics of light mass,high flexibility,slight damping and low natural frequency.As the span increases, the natural frequencies generally decrease,and the susceptibility of a roof structure with long-span to resonant excitation by turbulent wind action increases. Consequently,these structures have become progressively more wind sensitive. Therefore,the major objective of this research study is to further the understanding of wind effects on and structural behavior of long-span roof structures under strong wind actions by means of field measurements,wind tunnel tests and numerical prediction in order to apply such knowledge to design.This study includes five closely related parts.1.Field measurements of typhoon-generated characteristics near ground.Field measurements of typhoon-generated characteristics near ground are very useful, particularly for further understanding wind effects on long-span roof structures,and for incorporation into useable boundary layer wind flow simulation in wind tunnel tests. However,the chance to conduct field measurements of typhoon-generated characteristics near ground is quite rare,and obtained data are very important and valuable.Time series of wind speed and wind direction data were full recorded by two kinds of 3-D anemometers installed on the observation tower during the passages of Typhoon Chanchu(The maximum wind speed is 67.3 m/s) and Typhoon Prapiroon (The maximum wind speed is 40.7 m/s).The measured wind data are analyzed to obtain the information on mean wind speed and direction,turbulence intensity,gust factor,turbulence integral scale and probability distribution of fluctuating wind speed. This chapter presents some selected results including:(1) A 3-D anemometer with higher sampling frequency performs well in precision;(2) The probability density functions of fluctuating wind speeds approximately follow the normal distribution under lower wind velocity;(3) The gust factor is found to be linear with the longitudinal turbulence intensity.(4) The von-Karman and Kaimal type spectra are identified to be able to describe the energy distribution fairly well for the wind speed components in longitudinal direction of Typhoon Prapiroon and Typhoon Chanchu, respectively.2.Wind pressures on long-span roof structures.Investigations on the characteristics of wind loads and wind-induced response of long-span roofs have been made extensively.However,roof shapes of long-span structures vary widely from structure to structure.It is well known that wind effects on roof structures strongly depend on roof shape and incident wind flow characteristics.Consequently,it is difficult to propose a unified analytical approach to estimate wind loads and wind-induced response of various kinds of long-span roof structures.Therefore,there is a need to carry out comprehensive wind tunnel studies to further the understanding of wind effects on long-span roof structures.In this chapter,wind tunnel tests are conducted to investigate wind pressure distributions on two typical long-span roof structures under different wind directions;and the measured wind pressures,such as mean,root-mean-square(rms) and peak pressure coefficient distributions on the two roofs are presented and discussed.Furthermore,power spectra of fluctuating wind pressures measured from some typical taps located at the roof edges under different wind directions are presented.Based on these results,according to the characteristics of wind loads on the roofs,some roof configuration design strategies to improve the wind-resistance capacities of long-span roof structures are recommended.3.Prediction of wind-induced pressures on long-span roof structures using artificial neural networks.In wind tunnel experiments for long-span roof structures, it is usually necessary to install as more pressure taps as possible on model surfaces in order to capture the detailed characteristics of wind loads on the structures,since the cladding or roofing covers of the structures are very wind sensitive to the spatial variation and severe damages caused by wind-induced loading often occur in these locations.Although recent technological advances have made it possible to simultaneously measure surface pressures at more than 1000 locations on a building model,such experimental arrangement may still not be able to cover the whole surfaces of a large roof structure.Therefore,there is a need to explore an effective way to predict the wind-induced pressures on the entire roof structure on the basis of the pressure data from limited measurement points.The application of artificial neural networks(ANNs) to solve the title problem has received increasing interests in recent years.This chapter is concerned with developing two ANN approaches(a backpropagation neural network[BPNN]and a radial-basis function neural network [RBFNN]) for the prediction of mean,root-mean-square(rms) pressure coefficients, power spectra of fluctuating wind pressures and time series of wind-induced pressures on two typical long-span roof structures.Comparisons of the prediction results by the two ANN approaches and those from the wind tunnel test are made to examine the performance of the two ANN models,which demonstrates that the two ANN approaches can successfully predict the pressures on some regions of the large roof which are underwent small pressure variations on the basis of wind tunnel pressure measurements from a certain number of pressure taps.Meanwhile,it is also found that the prediction performance of the two ANN models for the interpolation case is better than that for the extrapolation case.Furthermore,an improved BPNN based a genetic algorithm(GA) is developed for the predictions of wind-induced pressures at some roof locations which are underwent large pressure variations with high rms pressure values due to strong flow separation;in which the number of weights is adjusted by GA to improve the convergence rapidity and stability of BPNN.It is shown through this chapter that the developed ANN approaches can be served as an effective tool for the design and analysis of wind effects on long-span roof structures in conjunction with wind tunnel tests.4.Wind-induced response of long-span roof structures.This chapter presents a new description of wind-induced response of long-span roof structures.It is noteworthy that in the proposed approach the total dynamic response is directly calculated by the complete quadratic combination(CQC) approach,in which the contributions of multimode response and modal response correlations are taken into consideration,and meanwhile the tapping influence coefficient approach is proposed to simplify the calculation of the node load vectors.Moreover,unlike existing approaches,it is not required to calculate the correlation of the load and the response, which is difficult to be determined by conventional methods.Finally,two typical extra-long-span roof structures are considered to illustrate the determination of the wind-induced response by the proposed approach and to demonstrate its effectiveness. On the other hand,special attention is also paid to the characteristics of wind pressures and wind-induced response of the spatial lattice structure and truss beam structure in this chapter.This chapter presents some selected resulting including:(1) wind-induced response and destroyed modes of the spatial lattic structure are very different from those of the truss beam structure;(2) the peak displacement response of the truss beam structure generally occurred in the middle of the span is induced by the holistic vibration of the roof;(3) the peak displacement response of the cantilevered spatial lattic structure generally occurred at the windward corner of the roof is induced by the partial vibration of the roof corner;(4) damping ratio and mode effects on power spectra of displacement responses of the nodes are analyzed;furthermore,the mechanism is also discussed in detail. 5.Full-scale measurements of wind effects on long-span roof structures. Although there have been many advances in wind tunnel testing and numerical simulation techniques for investigating wind effects on long-span roof structures, there are still many critical phenomena which can only be investigated by full-scale experiments.It has been widely recognized that the most reliable evaluations of dynamic characteristics and wind effects are obtained from experimental measurements of a prototype structure.In this chapter,full-scale measurements of wind effects on the long-span roof structure of Guangzhou International Exhibition Centre were conducted under strong wind action.Based on the field measurement results,a new method to identify the first several natural frequencies of the roof in vertical direction,using the dot matrix of power spectrum density approach,was presented.The new method eliminates the shortcoming of the conventional power spectrum density approach,such as the dummy apex and burr phenomena;and can perform well in precision.Comparison of the frequency results determined by the proposed method and those obtained from the finite element model analysis results was made to examine the applicability and accuracy of the proposed method.The field measurements can provide reliable but limited information.The wind tunnel tests can generate detailed and additional results that are not available from the field measurements.On the other hand,the ANN approach can be used as a supplement to wind tunnel tests to accurately estimate wind-induced pressures on long-span roof structures.Therefore,the field measurements,the wind tunnel tests and the ANN approach are complementary so that the understanding of wind effects on long-span roof structures can be improved.The outcome of this study is expected to be of considerable interest and practical use to professionals and researchers involved in the design of long-span roof structures.