考虑特征湍流影响的体育场悬挑屋盖脉动风压谱模型

基于风洞试验对体育场悬挑屋盖的脉动风压谱进行系统研究，旨在得到适用于此类结构的脉动风压谱模型，为风振响应分析提供必要的信息。通过对屋盖表面脉动风压进行谱分析，可知在屋盖前缘处的风压谱与来流风速谱较接近，但屋盖后缘处则差异很大，表现出明显的漩涡脱落特征。因此脉动风压自谱采用来流谱与漩涡脱落谱相结合的形式来描述，并通过权数因子体现屋盖表面不同位置处流场作用的特点。对于脉动风压互谱则用指数衰减函数来表示，并确定了适用于悬挑屋盖的衰减系数。为验证所提出风压谱模型的有效性及特征湍流对风致效应的影响，对系列悬挑屋盖结构进行风振响应分析，风荷载时程分别采用风洞试验测得的风压时程、基于建议风压谱模型模拟生成的风压时程、按拟定常假设生成的风压时程。基于建议模型得到的响应结果与试验结果基本一致，基于拟定常假设的风振响应极值偏小10%～15%，均方根值偏小30%～40%，脉动风压谱建模中不可忽略特征湍流的影响。

Spectral model of fluctuating wind pressure on grandstand roofs with consideration of signature turbulence

Based on wind tunnel tests of grandstand roofs with different pitches, the characteristics of fluctuating wind pressures were investigated, aiming to get available spectral model of fluctuating wind pressure, which was necessary in wind vibration analysis during simulating wind loading. The wind pressure spectra on roofs were obtained by spectral analysis. It is found that the shapes of wind pressure spectra of frontal taps are close to that of wind velocity spectra, but quite different from results of taps on the leeward part of roof which should be dominated by vortex shedding. The wind pressure spectra on roofs are modeled with the combination of approaching wind spectra and power spectrum representing vortex shedding, with the factors representing their contribution. The cross spectra between two points can be reached from an exponential function of the reduced frequency, the decay constant is decided from the wind tunnel data. In order to verify the availability of wind pressure spectra models on grandstand roofs and highlight the influence of signature turbulence, wind-induced vibration analysis in time domain for grandstand roofs with truss structure are carried out. The input time series are wind tunnel data, simulated wind pressure time series based on the spectral model of this paper and the wind pressure time series based on quasi-steady assumption respectively. The peak and fluctuating responses based on this paper’s model of wind pressure spectra are close to that of wind tunnel results, while the results based on quasi-steady assumption has maximum error of 10%-15% for peak responses, and 30%-40% for fluctuating responses. These results reveal that the effect of signature turbulence can not be ignored during modeling of fluctuating wind pressure spectra.