时频非平稳地震动模型及结构随机振动应用

地震动具有显著的时域和频域非平稳性,传统的均匀调制地震动模型不能充分表征时频两域的非平稳性,并影响结构随机振动分析的精准度。为了克服上述不足,建立由多峰平滑包络和非平稳噪声信号点乘而成的非均匀调制地震动模型,该模型能够在时域上充分反映地震动的波动和衰减特性,同时其频谱特征也与真实信号相接近。在此基础上,建立同时以时间和频率为变量的非平稳地震动演变谱随机模型。分析结果表明,以上两种模型在时域和频域的合成精度上均优于均匀调制模型。针对以上地震动随机模型的形式和特点,对传统结构随机振动响应求解方法进行改进。通过算例证明,按照均匀调制模型计算结构随机响应不能充分获得结构时频域的非平稳特性且结果偏于不安全,而利用两种改进模型可以获得更精确的结构响应概率特征和更细致的非平稳性演变过程。

Time-Frequency Nonstationarity Stochastic Ground Motion Model and Application in Structural Random Vibration

Ground motion involves significant nonstationary in time domain and frequency domain. The traditional uniform modulation model cannot fully characterize the nonstationary in both time domain and frequency domain, which debases the accuracy and precision of structural random vibration analysis. In order to overcome the above defects, a nonuniform modulated ground motion model is established, which is generated by the dot product of the multi-peak smooth intensity envelope and the nonstationary noise signal. The new model can fully represent the oscillation and attenuation characteristics of earthquake wave in time domain, at the same time, its spectral characteristics are close to the real values. Based on this study, the nonstationary evolution spectrum model is established, which simultaneously sets the time and frequency as variables. The analysis results show that these two models are superior to the uniform modulation model in both time domain and frequency domain. In view of the form and characteristics of the above seismic stochastic process model, the analytic method for solving the traditional structural vibration response is improved. Through a large number of examples, it is proved that the nonstationary characteristics in both time and frequency domains cannot be fully obtained when the structural random response is calculated according to the uniform modulation model, and the unsafe results are usually obtained. On the contrary, more accurate structural response characteristics and more detailed evolution process can be obtained when the two new improved models are used.