基于土层常规参数的液化发生概率计算公式及其可靠性研究

发布时间：2019-02-14 19:23

【摘要】：工程场地地震安全性评价和地震小区划工作在我国日益普及,对适于工程使用的土层液化发生概率计算方法的需求日趋强烈。该文沿用我国建筑抗震设计规范液化判别模式并使用同样基础数据,以地下水位、埋深、标准贯入击数等土层常规指标为直接变量,采用较为成熟的二分类Logistic回归分析理论,构造砂土液化概率计算公式和不同概率水平下液化临界值计算公式,通过回归分析和近期地震液化调查新数据分项检验其合理性和可行性。构造公式时采用我国大陆以往159例液化数据,检验公式时采用近期地震液化调查358例新数据,来源于1995年阪神地震和1999年集集地震。以上述两方面数据,检验概率50%的该文公式与现有建筑抗震设计规范确定性方法,结果表明:二者对国内大陆液化资料回判成功率基本相当,但该文公式略为保守;就阪神和集集地震新数据,该文公式不同砂层埋深检验结果均可接受,而现有规范砂层埋深下小于10m结果可以接受,大于10m时显著保守,方法需要改进。以上述两方面数据检验该文公式不同液化概率水平下的表现,结果表明:不同液化概率下的该文公式不仅定性上符合现有认识,而且非液化和液化概率水平相同时,对大陆非液化和液化场地回判成功率基本相当,对阪神和集集地震液化与非液化场地判别成功率基本相当,不同埋深下液化与非液化场地判别成功率大体一致。两方面数据检验表明,该文公式对不同液化概率水平、各种地震强度、地下水位和砂层埋深均有较好的适用性。
[Abstract]:Seismic safety evaluation and seismic microzonation of engineering sites are becoming more and more popular in China, and there is a growing demand for the probability calculation method of soil liquefaction which is suitable for engineering. In this paper, the liquefaction discriminant model of the code for seismic design of buildings in China is followed and the same basic data are used. The conventional soil layer indexes, such as groundwater level, buried depth and standard penetration number, are taken as direct variables, and the more mature theory of two-classification Logistic regression analysis is adopted. The calculation formula of liquefaction probability of structural sand and the formula of critical value of liquefaction at different probability levels are used to verify the rationality and feasibility of the formula by regression analysis and the new data of recent seismic liquefaction survey. The formula was constructed from 159 previous liquefaction data in mainland China and 358 new data from the recent earthquake liquefaction survey. The data were derived from the Hanshin earthquake in 1995 and the Jiji earthquake in 1999. Based on the above two data, the test probability of 50% of the formula and the existing building seismic design code certainty method, the results show that: the success rate of the domestic mainland liquefaction data judgment is basically the same, but the formula is a little conservative; For the new seismic data of Hanshin and Jiji, the test results of buried depth of different sand layers can be accepted in this paper, but the results of buried depth less than 10 m in existing codes are acceptable, and the method should be improved when the buried depth of sand bed is more than 10 m. The above two data are used to test the performance of the formula under different liquefaction probability levels. The results show that the formula under different liquefaction probability not only qualitatively accords with the existing knowledge, but also has the same non-liquefaction and liquefaction probability levels. The success rate of judging non-liquefaction and liquefaction sites in continental areas is basically the same, that of liquefaction and non-liquefaction sites in Hanshin and Jiji earthquakes is basically the same, and that of liquefaction and non-liquefaction sites under different burial depths is basically the same. Two aspects of data test show that the formula has good applicability to different liquefaction probability levels, various earthquake intensity, groundwater level and sand depth.
【作者单位】：中国地震局工程力学研究所;
【基金】：公益性地震行业科研专项(200708001) 国家自然科学基金(41272357,51208477) 国家科技部国际合作项目(2009DFA71720)
【分类号】：TU435

【参考文献】