地震动全非平稳时频谱统计模型及其在结构分析中的应用

Statistical model of time-frequency spectrum of earthquake and application in structural vibration analysis

  • 摘要: 针对传统的时域强度包络函数没有充分考虑地震动统计特性的不足,提出基于能量持时的地震动强度统计模型建立方法,获得不同场地类型的地震动强度包络统计模型及典型地震波。针对过滤白噪声型单峰型功率谱不能充分体现真实的地震动功率谱多峰性和强非平稳性的不足,采用更精准的多峰功率谱模型。在此基础上,基于小波包变换提出时变功率谱(演变谱)的统计模型和等效函数模型,二者均满足时频域的边缘条件及非负性,且可精准而细致地表征时频非平稳性。通过算例证明按照时变功率谱统计模型及其等效函数模型在结构进行随机响应分析时相比于传统的功率谱函数模型具有更高的精度。

     

    Abstract: Aiming at the insufficiency of the traditional intensity envelope function of time-domain does not fully represent the statistical characteristics of ground motions,a method of establishing statistical model of intensity envelope based on energy duration is proposed. The statistical model of intensity envelope of typical earthquake waves considering different site types is obtained. In view of the fact that the single peak power spectrum based on filtering white noise cannot precisely represent actual multi-peak phenomena and strong non-stationarity of power spectrum,a more accurate multi-peak power spectrum model is adopted. On this basis,a statistical model and an equivalent function model of time-varying power spectrum(evolutionary spectrum)are proposed based on wavelet packet transform. Both of them satisfy marginal condition and non-negativity in time-frequency domain,and characterize the non-stationarity in time-frequency domain accurately and intricately. The analysis examples show that the accuracy of the statistical model and its equivalent function model of time-varying power spectra are higher than the traditional power spectral function model for structural random vibration analysis.

     

/

返回文章
返回