In this paper the site categorization criteria and the corresponding site amplification factors proposed in the 2021 draft of Part 1 of Eurocode 8 (2021-draft, CEN/TC250/SC8 Working Draft N1017) are first introduced and compared with the current version of Eurocode 8, as well as with site amplification factors from recent empirical ground motion prediction equations. Afterwards, these values are checked by two approaches. First, a wide dataset of strong motion records is built, where recording stations are classified according to 2021-draft, and the spectral amplifications are empirically estimated computing the site-to-site residuals from regional and global ground motion models for reference rock conditions. Second, a comprehensive parametric numerical study of one-dimensional (1D) site amplification is carried out, based on randomly generated shear-wave velocity profiles, classified according to the new criteria. A reasonably good agreement is found by both approaches. The most relevant discrepancies occur for the shallow soft soil conditions (soil category E) that, owing to the complex interaction of shear wave velocity, soil deposit thickness and frequency range of the excitation, show the largest scatter both in terms of records and of 1D numerical simulations. Furthermore, 1D numerical simulations for soft soil conditions tend to provide lower site amplification factors than 2021-draft, as well as lower than the corresponding site-to-site residuals from records, because of higher impact of non-linear (NL) site effects in the simulations. A site-specific study on NL effects at three KiK-net stations with a significantly large amount of high-intensity recorded ground motions gives support to the 2021-draft NL reduction factors, although the very limited number of recording stations allowing such analysis prevents deriving more general implications. In the presence of such controversial arguments, it is reasonable that a standard should adopt a prudent solution, with a limited reduction of the site amplification factors to account for NL soil response, while leaving the possibility to carry out site-specific estimations of such factors when sufficient information is available to model the ground strain dependency of local soil properties.

本文首先介绍了欧洲法规8第1部分的2021年草案（2021草案，CEN / TC250 / SC8工作草案N1017）中提出的场地分类标准和相应的场地放大因子，并将其与当前版本的欧洲法规8进行了比较，以及最近经验性地面运动预测方程式中的站点放大因子。然后，通过两种方法检查这些值。首先，建立了一个强大的运动记录数据集，根据2021草案对记录站进行分类，并根据参考岩石条件从区域和全球地面运动模型计算站点到站点的残差，凭经验估算频谱的放大率。其次，对一维（1D）站点扩增进行了全面的参数数值研究，基于随机生成的剪切波速度剖面，并根据新标准进行分类。两种方法都可以找到一个相当好的协议。最相关的差异发生在浅层软土条件下（土壤类别E），由于剪切波速度，土壤沉积物厚度和激发频率范围之间的复杂相互作用，无论是记录还是一维散射都显示出最大的差异。数值模拟。此外，由于非线性（NL）场地效应的影响较大，针对软土条件的一维数值模拟倾向于提供比2021草案更低的场地放大因子，并且低于记录中相应的场地间残差。在模拟中。对三个KiK-net站点的NL效应进行了针对特定地点的研究，并记录了大量的高强度地面运动，为2021草案的NL减小因子提供了支持，尽管允许进行此类分析的记录站点数量非常有限，但无法推导更一般的含义。在存在此类争议性论点的情况下，标准应采用审慎的解决方案，并合理减少场地放大因子，以解决NL土壤反应，这是合理的，同时保留了针对此类因素进行特定于场地的估计的可能性。当有足够的信息可用来模拟当地土壤特性对地面应变的依赖性时。尽管允许进行这种分析的记录站数量非常有限，但仍无法得出更一般的含义。在存在此类争议性论点的情况下，标准应采用审慎的解决方案，并合理减少场地放大因子，以解决NL土壤反应，这是合理的，同时保留了针对此类因素进行特定于场地的估计的可能性。当有足够的信息可用来模拟当地土壤特性对地面应变的依赖性时。尽管允许进行这种分析的记录站数量非常有限，但仍无法得出更一般的含义。在存在此类争议性论点的情况下，标准应采用审慎的解决方案，并合理减少场地放大因子，以解决NL土壤反应，这是合理的，同时保留了针对此类因素进行特定于场地的估计的可能性。当有足够的信息可用来模拟当地土壤特性对地面应变的依赖性时。