【作者】 耿淑伟；

【导师】 陶夏新；

【作者基本信息】 中国地震局工程力学研究所 ， 防灾减灾工程及防护工程， 2005， 博士

【摘要】 抗震设计规范是抗震设计的技术性法规,以现有的科学水平和经济条件为前提、以现有的经验和资料为依据制定。每个版本的抗震设计规范中,对地震作用的规定都是重要的一章,是抗震设计验算和采取抗震构造措施的基础,也是抗震设计规范与其他设计规范最重要的本质差别。本文在系统总结回顾国内外抗震设计规范中的设计地震作用规定的发展、演变过程的基础上,以大量实际强震记录为基础,对地震作用规定的五个方面进行了系统的研究,提出了改进的建议,主要包括: 1 本文根据254 个美国西部有强震记录的台站的剪切波速度结构资料,按中国建筑抗震设计规范的场地划分标准,对这些台站进行了场地分类,得到了有明确场地类别资料的、在自由地表或建筑物(低于三层的建筑物)地下室或一层取得的713 条记录。为了研究地震动数字记录的反应谱长周期段可靠性,本文还专门选用了集集地震中获得的预存储10 秒以上的、质量较好的记录132 条;为研究了长周期反应谱的形状,另整理了质量较好的C、D、E 类场地的402 条记录。这些基础数据,与从相关研究成果(王国新,2000)直接引用的1220 条强震记录,为本文的研究工作提供了坚实的基础。 2 研究了强地震动记录在长周期段的可靠性。将强地震动观测数据分为模拟记录和数字记录两大类。对于模拟记录,其加速度反应谱在5-6 秒以内是可靠的;集集地震中获取的数字记录对比结果显示,在周期小于10.24 秒范围内,信/噪谱比均在90 倍以上,表明在周期10 秒以内,噪声对反应谱影响很小。比较通过EMD 调整基线的加速度反应谱与常规数据处理得到的反应谱,指出基线漂移对加速度反应谱的影响可忽略不计。 3 在讨论地震动加速度反应谱幅值随周期下降的机理和主要特征的基础上,针对模拟记录和数字记录的质量差别,分别研究了6 秒以内和更长周期段反应谱下降的形状,指出GB50011-2001 规定的方案过于保守;通过对比数字记录和模拟记录长周期段的特征,论证了模拟记录的统计特征向10更多还原

【Abstract】 Seismic design code is a technique rule for seismic design. Ii is established relaying on current experiences and data and matches science and economical development level. The provision of earthquake action is the most important chapter in various versions of the seismic design code. It is the foundation for seismic verifying and seismic reinforce. And it shows the essential difference between the Seismic design code and other design codes. After evolutive process of earthquake action provision in main countries is summarized, 5 aspects of earthquake action provision area studied on the basis of abundant strong motion records, suggestions for improvements are put forward. The main achievements in this study could be summarized as follows: 1 Shear wave velocity profile data at 254 stations with strong motion records are collected in the western part of the USA. The stations are classified by site condition from the provisions of Code for Seismic Design of Buildings (GB50011-2001). 713 records are also gained on free field, in 1st floor or basement in low (no more than 3 stories) buildings. 132 digital records with longer than 10 seconds noise pre-record are collected in Taiwan ChiChi earthquake to study the reliability of long period response spectra. 402 high-quality records on C, D and E site type are collected for the study of response spectra shape in long period. The records mentioned above and 1220 strong motion records quoted from the results (Wang G. X., 2000) offer a strong foundation for the study. 2 The reliable range of response spectra in long period is studied. The records are classified as analog and digital ones. For the former, the response spectra are reliable in the period range less than 6 second. For latter, in ChiChi earthquake, the long period range is less than 10 second, the ratios between the Fourier spectra from records and the pre-recorded noise are bigger than 90, the fact shows that the effect of noise to record can be ignored. The effect of baseline offset to acceleration response spectra at long period is examined by comparing the response spectra from baseline corrections by EMD and the general ones. The result shows that the effect can be ignored. 3 On the foundation of the study of mechanism and main character about amplitude of response spectra decreasing with period increasing, the response spectra shape in the long period is studied for both analog records and digital records. The result shows that the provision on response spectrum shape in the code of GB50011-2001 is very conservative. The comparison of the response spectra character in long period between analog and digital records show it is feasible to extend response spectra from analog records to longer period. It is pointed out that the response spectrum curve in 6 second cad be extend to 10 second with descent rate of T-1 and the lower limitation must be deleted. 4 That current research and development tendency is summarized shows that it is necessary to provide the maximum amplitude of response spectrum with considering site condition too. The difficulty and emphasis are pointed by comparing the existing research. A direct comparison method in the study is developed. The statistical result from strong motion records shows that the maximum amplitudes of response spectra are varied with site category. The relation of the variety with site classifying criterion and earthquake intensity is built. Referencing provision in 1997 NEHRP and other research, the proposed site coefficients are put forward based on statistical result in the study. In the suggestion, the maximum amplitude of response spectrum is determined from the one in the code of GB50011-2001 multiplying by the proposed site coefficient. The improvement is not only in short period but also in long period. 5 The ratios of vertical to horizontal response spectra are calculated from abundant records. The essential character that ratios varied with period is emphasized. It is found that in general, there is a peak in short period on the ratio curve, then a sag, and a constant in long period. The statistical result shows that the peak is not high, and the sag is shallow on site category I, so that the ratio can be considered as a constant independent of period. The peak is near 1.0, and the period of sag is 0.5 second or so, the constant is just a little more than 0.5 on site category II, III and IV, so it is suggested that the vertical response spectra could be determined by horizontal one multiplying by a spectrum ratio function dependent of period. The simplified spectra ratio function is developed from the analysis of the effect of magnitude, epicenter distance and site category on the ratios. For the analysis of the physics essence of V/H spectra ratio, the character of Fourier spectra ratios are analyzed qualitatively when p-wave, s-wave and Rayleigh wave radiate on different site models. It is found that the soil deposits velocity profile is the control factor to Fourier spectra ratios onthe ground surface. 6 82 strong motion records in rock site are adopted to simulate the envelope curve of ground motion time history. The parameters of the envelope curve are picked up. The PSDF of every stationary random process is calculated. They are used to simulate standard spectrum density function, and pick-up the parameters of PSDF. The simplified method are developed to determine the envelop function parameters andβg、ωc value in PSDF. The attenuation functions of S0 和ωg are fitted. The response spectra parameters from the conversion from the PSDF in the study are compared with the ones directly from attenuation function. It is found that character frequencies and the amplitudes of response spectra are comparative from the two methods. So the suggestion are put forward that the PSDF can be determined from the one converted from design response spectra.更多还原