【摘要】 人工合成地震动的随机性对地震响应结果有重要影响，足够的输入地震动时程数量可以确保结果的稳定性和准确性。为研究地震动随机性对黄土场地地震响应的影响并确定基岩输入地震动时程的必要数量，基于黄土地区场地土层剪切波速和非线性参数资料，建立了4个不同覆盖层厚度的土层地震反应分析模型，以地震危险性分析得到的黄土地区基岩加速度反应谱为目标谱，人工合成地震动过程中随机数取值为1～9 801,合成9 801条随机地震动时程作为土层地震反应输入，通过土层地震反应分析得到地表响应结果。对比分析了不同工况下，随机数对黄土场地地表峰值加速度和加速度反应谱的分布特征及其差异，结果表明：地表峰值加速度整体上遵循正态分布的特征，变化范围约为均值的0.77～1.71倍；加速度反应谱主要在周期小于0.1 s的高频部分受到影响，而整体形状所受影响较小。通过统计学方法给出，不同工况下黄土场地基岩输入地震动加速度时程的必要数量，在各工况下，所需的必要时程数量最大为12条，最小为4条。更多还原

【Abstract】 The randomness inherent in artificially synthesized earthquake motions significantly influences the outcomes of seismic response analyses. A sufficient quantity of input earthquake motion time histories is essential to ensure the stability and accuracy of the results. This paper aims to investigate the impact of earthquake motion randomness on the seismic response of loess sites and to ascertain the requisite number of rock input earthquake motion time histories. To achieve this, four distinct layered seismic response analysis models are established based on shear wave velocity and non-linear parameter data from the loess region. The target spectrum is derived from the earthquake risk analysis of the bedrock acceleration response spectrum in the loess area. Random numbers utilized during the artificial synthesis of seismic motion range from 1 to 9 801, resulting in the synthesis of 9 801 random seismic motion time histories for input into the layered seismic response analysis. The surface response results are obtained through this layered seismic response analysis. A comparative analysis of varying working conditions reveals the distribution characteristics and differences of random numbers concerning the peak ground acceleration and acceleration response spectrum of the loess site. The findings indicate that the peak ground acceleration generally conforms to a normal distribution, exhibiting a variation range of approximately 0.77 to 1.71 times the mean value. Furthermore, the acceleration response spectrum is predominantly influenced in the high-frequency range, specifically with periods of less than 0.1 s, while its overall shape remains relatively unaffected. Employing statistical methods, this study presents the necessary number of input earthquake motion acceleration time histories for the bedrock of loess sites under different working conditions, with the required number ranging from a maximum of 12 to a minimum of 4 across various conditions.更多还原