【作者】 王栋；

【导师】 谢礼立；

【作者基本信息】 中国地震局工程力学研究所 ， 岩土工程， 2010， 博士

【摘要】 近断层地震动由于其自身不同于远场地震动的特征以及对工程结构产生的严重破坏,成为近年来工程地震界的研究热点。近断层地震动的影响因素众多,除受常规的震源、路径以及场地条件影响外,还与上/下盘效应和破裂方向性效应密切相关,两者对近断层地震动的峰值、频谱、持时以及空间分布产生重要的影响。尤其是发生在倾斜断层上的地震,比如北岭地震、集集地震以及汶川地震,上/下盘效应对近断层地震动的影响更加突出,位于断层两侧即上下盘的近断层地震动存在较为明显的差别。因此充分认识近断层地震动上/下盘效应,对于因地制宜的采取工程结构的抗震措施具有重要意义。以往研究虽然对上/下盘效应有所涉及,但还不够系统深入。本文全面细致的比较了近断层上下盘地震动的差别,即近断层地震动上/下盘效应的特征、产生的原因以及影响因素。通过对近断层地震动的数值模拟和实际地震的强震数据分析两个方面,从地震动的峰值、频谱以及持时三个角度系统的分析了加速度、速度以及位移的三分量的上/下盘效应的基本特征,并从本质上揭示了上/下盘效应产生的原因,深入具体地剖析了上/下盘效应的影响因素。本文研究的主要内容和结论如下:1.基于北岭地震和集集地震的强震记录,系统地分析了上/下盘效应对近断层地震动峰值、频谱和持时的影响。研究表明:①断层距在30公里以内的上盘地震动峰值(加速度峰值PGA、速度峰值PGV和位移峰值PGD)大于该距离范围内地震动峰值的平均水平,上/下盘效应对PGA、PGV及PGD的影响依次减弱。②断层距在30公里以内的上盘反应谱值(加速度、速度以及位移反应谱值)在较宽的周期范围上(周期小于4s)明显大于该距离范围内的反应谱值的平均水平。③断层距在30公里以内的上盘地震动持时略小于该距离范围内地震动持时的平均水平。④上/下盘效应对平行断层走向分量(Fault Parallel,FP)、垂直断层走向分量(Fault Normal,FN)和竖向分量(UP Down,UD)地震动的影响依次减弱。2.比较了常见的观测点到断层面的距离标准,引入了均方根距离的定义,分析了上/下盘效应产生的原因。研究表明:①具有加权平均意义的均方根距离不同于常见的断层距、Campbell距离、震源距和震中距等距离标准,可以准确地描述观测点到断层面的整体靠近程度及倾斜断层的不对称程度。②使用均方根距离对地震动参数进行回归时,近断层地震动的上/下盘效应不明显,可以忽略其影响。③近断层地震动的上/下盘效应是由于具有相同断层距的上下盘观测点,上盘观测点在整体上距离断层破裂面较近即上盘观测点的均方根距离较小引起的,即上/下盘效应是由倾斜断层的不对称分布引起的几何效应。3.数值模拟了不同断层模型下逆断层型地震近断层地震动的上/下盘效应,分析了各种震源参数包括断层倾角、断层的上界埋深、断层的破裂速度、断层的破裂初始位置以及震级对上/下盘效应的影响。以下分别进行说明:(1)对于倾斜的逆断层型地震,上/下盘效应对地震动峰值、反应谱和持时影响的基本特征:①上盘地震动峰值(PGA、PGV及PGD)大于具有相同断层距的下盘观测点的地震动峰值。上/下盘效应对UD、FP和FN分量的影响依次减弱,上/下盘效应对PGA,PGV和PGD的影响依次增强,影响的强弱表现在两个方面:一是上、下盘地震动峰值差别的大小;二是上下盘峰值比大于1的区域(即上/下盘效应影响区域)的大小。②上盘三分量加速度反应谱值明显大于相同断层距下盘观测点的反应谱值。上/下盘效应对长周期的加速度反应谱值影响大于短周期;上/下盘效应对三分量的影响程度因周期值的不同而不同。③上盘三分量加速度持时明显小于相同断层距的下盘观测点的加速度持时,随着断层距的增大上下盘加速度持时的差别先增大后保持不变。上/下盘效应对三分量加速度持时的影响大致相当。(2)断层的上界埋深对上/下盘效应的影响:①上界埋深对地震动峰值的上/下盘效应的影响因地震动分量和PGA,PGV和PGD的不同而不同,但是总体来说,上/下盘效应对地震动峰值的影响随着上界埋深的增大而增强,表现为上下盘地震动峰值比增大且上/下盘效应影响区域向远离断层破裂迹线的方向扩展。②上界埋深对反应谱值的上/下盘效应的影响因地震动分量和周期值的不同而不同,但是总体来说,上/下盘效应对加速度反应谱值的影响随上界埋深的增大而增强,埋深的变化对短周期加速度反应谱值的上/下盘效应影响比长周期强。③断层上界埋深对加速度持时的上/下盘效应的影响因断层距的不同而不同,在靠近断层区域(断层距小于20km)随着上界埋深的增大上下盘持时的差别减小,而在远离断层的区域随着断层上界埋深的增大上下盘地震动持时差别增大。(3)断层倾角对上/下盘效应的影响:①断层倾角对地震动峰值的上/下盘效应的影响因地震动分量和PGA,PGV和PGD的不同而不同,但是总体来说上/下盘效应对地震动峰值的影响随着断层倾角的减小逐渐增强,表现为上下盘地震动峰值比增大且影响区域向远离断层破裂迹线的方向扩展。②断层倾角对反应谱值的上/下盘效应的影响因地震动分量和周期值的不同而不同,但是总体来说,上/下盘效应对加速度反应谱值的影响随着断层倾角的减小逐渐增强,断层倾角的变化对短周期加速度反应谱的上/下盘效应影响比长周期强。③随断层倾角的增大上下盘加速度持时的差别减小。(4)断层的破裂速度对上/下盘效应的影响:①破裂速度对地震动峰值的上/下盘效应的影响因地震动分量和PGA,PGV和PGD的不同而不同,总体来说随着破裂速度的增大,上/下盘效应对地震动峰值的影响略有增强,表现在上下盘地震动峰值比增大。②破裂速度对加速度反应谱值的上/下盘效应的影响因地震动分量和周期值的不同而不同,总体来说随着破裂速度的增大,上/下盘效应对加速度反应谱值的影响略有增强,表现在上下盘加速度反应谱值比增大。③随着断层破裂速度的增大上下盘加速度持时的差别增大。④在超剪切破裂状态下(破裂速度大于剪切波速)的PGA的上/下盘效应比亚剪切破裂状态下的明显,而PGV和PGD的上/下盘效应不如亚剪切破裂状态(破裂速度小于剪切波速)下的明显;超剪切破裂状态下的短周期加速度反应谱值的上/下盘效应比亚剪切破裂下的明显,而长周期加速度反应谱值的上/下盘效应不如亚剪切破裂状态下的明显。(5)断层的初始破裂位置对上/下盘效应的影响:破裂起始点沿平行断层走向方向的位置变化对近断层地震动的上/下盘效应的影响不明显,而其沿断层下倾方向的位置变化对上/下盘效应的影响显著,主要表现在:①上下盘地震动峰值的差别当破裂起始点位于断层上界时最大,位于断层下界时次之,位于断层下倾方向中心时最小。②上下盘短周期加速度反应谱值的差别随破裂起始点的变化规律与地震动峰值相同;上下盘长周期加速度反应谱值的差别随着破裂起始点由断层上界向下界的移动而逐渐减小,而且上/下盘效应的影响区域向远离断层破裂迹线的方向扩展。③随着破裂初始点由断层上界向下界的移动,上下盘加速度持时的差别逐渐增大。(6)震级(断层宽度)对上/下盘效应的影响:①震级对地震动峰值的上/下盘效应的影响因地震动分量和PGA,PGV和PGD的不同而不同,总体来说震级的增大使得上下盘地震动峰值差别增大,且上/下盘效应的影响区域向远离断层破裂迹线的方向扩展。②震级对加速度反应谱值的上/下盘效应的影响因地震动分量和周期值的不同而不同,随着震级的增大上下盘加速度反应谱值比增大,上/下盘效应的影响区域向远离断层破裂迹线的方向扩展。③随着震级的增大,上下盘地震动持时差别增大。(7)利用均方根距离对比了上/下盘效应的影响因素,发现影响断层不对称分布因素如断层上界埋深、断层倾角以及地震震级也是上/下盘效应的主要影响因素,进一步验证了上下盘效应是由倾斜断层的不对称分布引起的几何效应。4.数值模拟了倾斜的走滑型和正断层型地震的上/下盘效应,并研究了断层类型对上/下盘效应的影响。研究表明:①走滑型地震,地震动峰值和加速度反应谱值的上/下盘效应表现在UD,FN和FP三个分量;逆断层型和正断层型地震,上/下盘效应主要表现在FP和UD两个分量。②对于地震动峰值和加速度反应谱值的上/下盘效应,走滑型地震最为明显,其次为逆断层型地震,再次为正断层型地震。表现在上下盘地震动峰值比和加速度反应谱值比以及上/下盘效应影响区域的大小按照走滑、逆冲和正断层型地震的顺序依次减小。③对于地震动持时的上/下盘效应,正断层和逆断层型地震比走滑型地震明显。5.利用汶川地震的强震观测记录,分析和总结了汶川地震近断层地震动的上/下盘效应特征,并比较了汶川地震的强地震动参数与国内外现有衰减模型的差别。研究表明:①汶川地震的近断层地震动上/下盘效应对地震动峰值的影响主要体现在加速度峰值PGA,而对速度峰值PGV和位移峰值PGD的影响不明显。②汶川地震的近断层地震动的上/下盘效应对反应谱值的影响表现在短周期,即周期小于0.5秒的加速度反应谱值,而对长周期加速度反应谱值影响不明显。③汶川地震的近断层地震动的上/下盘效应对地震动持时的影响主要表现在加速度持时,而对速度和位移持时的影响不明显。④与国内外地震动参数的衰减模型相比,汶川地震的短周期(周期小于0.5s)地震动明显大于预测值,而长周期(周期大于1s)地震动明显小于预测值。更多还原

【Abstract】 Study on the near-fault ground motion has been the hot research topic in the engineering seismological community for more than 10 years, which results from the near-fault ground motion having the different characteristics from far-fault ground motion and its serious damage on the structures. Beside the conventional influence factors on the ground motions, such as source, path and the site condition, both the hanging wall/foot wall effects (the HW/FW effects) and rupture directivity have great influences on the near-fault ground motions. For the large earthquakes occurred on non-vertical faults, such as the Northridge earthquake, Chi-Chi earthquake and Wenchuan earthquake, there are obvious differences between the ground motions on the hanging wall side and those on the foot wall side. This study focuses on the ground motion differences between the hanging wall and foot wall, i.e., the HW/FW effects on the near-fault ground motions. Three aspects involved are mainly studied, the characteristics, influence factors and reasons of the HW/FW effects. To deal with the three issues, we conduct analyses of both the real near-fault recordings and simulated ground motions. Through comparing the peak values, response spectra and durations of near-fault ground motions between the hanging wall and foot wall, some conclusions can be drawn:1. Analyses of the observed ground motions from the Chi-Chi and Northridge earthquakes indicate that the HW/FW effects have important influences on the peak values, response spectra and durations of ground motions. First, the three-component peak ground accelerations (PGA), peak ground velocities (PGV) and peak ground displacements (PGD) on hanging wall are larger than those on foot wall with the same rupture distances. Second, the tri-component spectral ordinates of accelerations at almost all period ranges are larger on the hanging wall than on the foot wall at the similar rupture distances. Third, the tri-component durations of ground motions on hanging wall are smaller than those on the foot wall. Last, the HW/FW effects have higher influences on horizontal component than the vertical component of near-fault ground motions, and the HW/FW effects on peak values of ground motions becomes lower and lower in terms of PGA, PGV and PGD.2. The commonly used site-to-source distance measures, such as rupture distance Drup, epicentral distance Depc, hypocentral distance Dhyp, Joyner-Boore distance DJB and seismological distance Dseis can not describe the general proximity of a site to the rupture plane. The root-mean-square distance Drms with the meaning of weighted-average is introduced to accurately represent the proximity between an observer and the rupture plane, and examine the HW/FW effects on the near-fault ground motions. The result shows that there is no obvious bias towards positive for the hanging wall residuals when the Drms is used as distance measure, which is opposite to the results when other distance measures are used. The HW/FW effects are demonstrated to be a geometric effects caused by the asymmetry of dipping fault.3. By numerical simulation of strong ground motions from different thrust-fault events, the influence factors of the HW/FW effects are studied, such as the depth to the top of rupture (Ztop), the fault-dip (Dip), the rupture velocity (Vr), the location of initial point of rupture and magnitude.(1)Variation of Ztop has an obvious effect on the ground motions differences between hanging wall and footwall.①With increasing of Ztop, the ratio between hanging wall and footwall (HW/FW ratio) for amplitudes (PGA, PGV and PGD) of ground motions becomes larger and the affected area of the HW/FW effects extends greatly. The variation of Ztop does not change the HW/FW effects on FN component, but UD and FP components of peak values.②The influence of Ztop on the HW/FW effects on response spectra acceleration (RSA) depends on component and period. For short period, the UD- and FP- components HW/FW ratio for RSA become larger and affected area extends if Ztop increases. For long-period, the UD-component HW/FW ratio decreases but the FP-component HW/FW ratio increases with the increasing of Ztop.③With the increase of Ztop, the difference of the duration between hanging wall and footwall reduces in the near-fault region.(2)Variation of Dip has a great effect on the ground motion differences between hanging wall and footwall.①For the amplitudes of ground motion, the HW/FW ratios for PGA, PGV and PGD increase with the decreasing of Dip.②For the RSA, the variation of Dip on the HW/FW effects depends on component and period. However, in general, the HW/FW ratios for three-component RSA increase and the affected areas of HW/FW effects extend greatly if Dip decreases.③For the duration of ground motions, the difference of duration between HW and FW enlarges when Dip reduces.(3) Variation of Vr has an effect on the ground motion differences between hanging wall and footwall.①For the amplitudes of ground motions, the HW/FW ratios for PGA, PGV and PGD decrease with the increase of Vr②For the RSA, the variation of Vr on the HW/FW effects depends on component and period. However, generally speaking, the HW/FW ratios for RSA decrease when Vr increases, but the affected area of HW/FW effects does not change a lot.③For the duration of ground motions, the difference of duration between hanging wall and footwall enlarges with the increase of Vr.④The HW/FW effects on PGA are more obvious under super-shear rupture(The velocity of rupture is larger than the velocity of S wave) status than under sub-shear rupture(The velocity of rupture is smaller than the velocity of S wave) status, however, the HW/FW effects on PGV and PGD are opposite. The HW/FW effects on RSA at short-period are more obvious under super-shear rupture status than under sub-shear rupture statu, however, the HW/FW effects on RSA at long-period are opposite.(4) The location of initial point of rupture has an effect on the ground motions difference between HW and FW. In general, the variation of location for initial point of rupture along fault-parallel direction does not influence the HW/FW effects, but the variation along the down-dip direction changes the HW/FW effects a lot.①For the amplitude of ground motions, with the initial point ascending from the bottom of the rupture, to the center of down-dip direction, and then the top of rupture, the HW/FW ratios of PGA, PGV and PGD increases and the affected area of HW/FW effects extends.②For the short-period RSA, with the initial point changing from the top of rupture, to the bottom of the rupture, and then the center of down-dip direction, the HW/FW ratio for RSA decreases. However, for the long-period RSA, the HW/FW effects become more and more obvious when the initial point changes from the bottom of the rupture, to the center of down-dip direction, and then the top of rupture.③For the duration of ground motion, the difference of duration between HW and FW monotonously enlarges with the initial point changing from the top of the rupture, to the center of down-dip direction, and then the bottom of rupture.(5) Variation of Magnitude has an effect on the ground motion differences between hanging wall and footwall.①For the amplitudes of ground motion, the HW/FW ratios for PGA, PGV and PGD increase with the increasing of Magnitude, but the affected areas of HW/FW effects extend.②For the RSA, the variation of Magnitude on the HW/FW effects depends on component and period. However, in general, the HW/FW ratios for three-component RSA increase if Magnitude increases, but the affected areas of HW/FW effects extend.③For the duration of ground motions, the difference of duration between HW and FW reduces when Magnitude reduces.(6)Using root-mean-square distance, we compare the influencing factors of the HW/FW effect. The result indicates the factors which greatly change the asysmetry of dipping fault such as depth to the top of fault, fault dip and magnitude are also the major influencing factors of HW/FW effects.4. By simulating the ground motions from the non-vertical strike-slip and normal events, the hanging wall/footwall effects on near-fault ground motions are examined. The focal mechanism on the hanging wall/footwall effects is also studied. The result show:①The HW/FW effects on near-fault ground motions also exist in the strike-slip event as long as the earthquake occurs on non-vertical fault.②The HW/FW effects on near-fault ground motions also exist in the normal event as long as the earthquake occurs on non-vertical fault.③For the normal and thrust event, the HW/FW effects mainly embody on UD and FP component of ground motion. However, for the strike-slip event, the HW/FW effects embody on three components of ground motion④For the strike-slip, normal and thrust event with the same earthquake magnitude and the same fault dip, the HW/FW effects on near-fault ground motion are most obvious in strike-slip event, and least obvious in the normal event.5. Based on the observed ground motions from Wenchuan earthquake, the hanging wall/footwall effects on near-fault ground motions are studied, and the comparisons between the observed motions from this event and the predicted motions from contemporary attenuation relations are also conducted.①The HW/FW effects on amplitudes of ground motion mainly embodies on PGA.②The HW/FW effects on response spectra accelerations mainly embody on the short-period (T≤0.5s) ground motions.③The HW/FW effects on duration mainly embodies on the duration of acceleration.④The observed short-period (T≤0.5s) ground motions from the Wenchuan earthquake are under-predicted, however, the long-period (T>1s) ground motions are greatly over-estimated by the contemporary attenuation relations world-wide.更多还原