场地条件校正的PGA放大系数简化估计方法

Simplified Prediction Method for PGA Amplification Factors Corrected by Site Conditions

地表峰值加速度(peak ground-motion acceleration,PGA)是直观反映地震动强度的一个物理量,概念清晰且工程应用方便.场地条件校正的PGA及其校正方法是特定工程抗震设计需要解决的问题.为此,选取日本具有地表和井下记录KiK-net台网的32个台站及地震数据,通过对实测地震数据分析,提出了一种场地校正的PGA放大系数(f_(PGA))的简化估计方法.该方法通过数据回归给出f_(PGA)概率密度函数参数与场地特征参数线性组合之间的线性和二次函数形式的拟合公式;并采用f_(PGA)概率预测模型,提出了多超越概率水平地表PGA值的预测方法.数据分析结果显示:场地f_(PGA)具有随机不确定性,可以采用对数正态分布函数模拟,其概率密度函数的参数(均值和标准差)与单一场地特征参数相关性较小,但与场地特征参数的线性组合相关性较大.模型预测值与实测数据吻合较好,验证了简化估计方法的可行性.

Peak ground-motion acceleration(PGA)directly reflects ground shaking intensity with the advantages in conceptual clarity and engineering application.Prediction of PGA,along with site condition correction,needs to be handled in site-specific seismic design.In this work,32 stations and the earthquake data recorded are collected from the KiK-net strong motion array,Japan,so as to propose a simplified method to predict PGA amplification factors(f_(PGA))corrected by site condition.Linear and quadratic empirical formulae of the f_(PGA) possibility model parameters with respect to the combinations of site characteristic parameters are obtained via regression analysis.Using the f_(PGA) model,ground surface PGA values,corrected by site conditions,can be predicted under different exceedance probability levels.Data analysis indicates that f_(PGA) is variable but can be simulated by a log-normally distributed function,of which mean and standard deviation are less correlated with a single site characteristic parameter but have good correlation with the linear combinations of the site characteristic parameters.The reasonable agreement between the predictions and records testifies the feasibility of the proposed method.