考虑随机效应的Clough-Penzien功率谱参数预测模型及地震动合成

Parameter prediction model of Clough-Penzien power spectrums considering random effects and ground motion synthesis

国内外学者对典型的地震动参数(a PG,v PG,SAs,I A,Ds等)预测方程的研究不断深入,但对于表征地震动频谱特性的功率谱模型参数预测方程的研究却很少。鉴于地震动的频谱特性与震源特性、传播路径和场地条件紧密相关,该研究基于地震动传播的物理机制建立了Clough-Penizien功率谱(C-P谱)模型参数的预测模型。为了更好地描述不同地震事件不同记录之间的相关性,对模型考虑随机效应,建立了谱参数的预测随机效应模型。采用改进的随机效应一次迭代回归算法,对各类场地C-P谱模型参数的预测随机效应模型进行回归分析,确定了模型系数。结合相位谱的物理传播模型与所提出的C-P谱参数预测模型,构建了“震源-传播路径-局部场地”物理传播机制的随机函数模型。算例分析结果表明,该研究提出的谱参数预测方程模型与实际功率谱拟合较好,模拟得到的地震动记录与原始地震记录的时频特性十分接近,可以较好实现地震动记录的模拟和预测,对地震危险性、易损性和风险分析及抗震设计与评定具有重要的指导意义。

There have existed many investigations on the prediction equations of typical ground motion parameters,such as a PG,v PG,SAs,I A,Ds,etc.However,there has been little research on the prediction equations of power spectrum model parameters,which represent ground motion frequency-spectrum characteristics.In view of the close relationships between the spectral characteristics of ground motions and the source characteristics,propagation path and site conditions,a prediction model of Clough-Penzien power spectrum parameters was established based on the physical mechanism of ground motion propagation.In order to better describe the correlation between different records of different seismic events,considering the random effect of the model,the random effect prediction model of spectral parameters was established.The improved one-time iterative regression algorithm for random effects was used to conduct the regression analysis on the random effect prediction models of various parameters of the Clough-Penzien power spectrum at various sites,and the model coefficients were determined.Finally,combining the physical propagation model of phase spectrum with the proposed prediction model of Clough-Penzien power spectrum parameters,a random function model of the“source propagation-site”physical propagation mechanism was constructed.It is shown by a practical example that the predicted spectral parameters fit well with the actual power spectrum parameters,and the time-frequency characteristics of the simulated seismic records are very close to those of the original seismic records,so the method proposed is helpful to the analyses of seismic hazard,fragility and risk as well as to the seismic design and evaluation of engineering structures.