基于逻辑回归的砂土液化判别研究

发布时间：2018-03-11 06:10

本文选题：地震　切入点：砂土液化　出处：《中国地震局工程力学研究所》2017年硕士论文　论文类型：学位论文

【摘要】：砂土液化是一种典型的地震灾害现象,液化判别的概率表达是岩土地震工程研究的热点问题之一,深入的探索砂土液化机理和发展考虑不确定性的液化判别方法,对工程场地条件的评价具有重要的理论意义的工程应用价值。本文以我国大陆地区已有的液化调查数据为基础,利用逻辑回归分析方法建立了砂土液化概率判别公式和给定液化概率的临界标准贯入击数的计算公式。通过对我国大陆地区和台湾集集大地震液化调查数据的回判和预判分析,并对比我国现行建筑抗震设计规范的判别结果,对本文给出的液化概率判别公式的适用性和判别结果的可靠性进行了检验。本文主要工作及研究成果如下:1.综述了国内外砂土液化的研究现状。总结了砂土液化的震害现象,介绍了砂土液化机理和液化影响因素,对国内外砂土液化判别方法的研究现状进行了总结和评述,并对各判别方法的优缺点及其适用性进行了简要述评,指出了现有砂土液化判别方法存在的问题与不足,并针对这些问题进行了讨论。2.收集整理和分析了我国大陆地区已有的液化调查数据。介绍了我国大陆地区发生8次大地震的背景资料,收集整理了大陆地区已有的液化调查数据,并对现有液化数据的烈度、标准贯入击数、砂层埋深和地下水位的分布特征进行了系统分析,为液化概率判别方法的建立提供了基础资料。3.利用逻辑回归建立了砂土液化概率判别公式。介绍了逻辑回归模型的基本原理和应用领域,选取了烈度、标贯击数、砂层埋深和地下水位4个参数作为液化评价指标,以我国大陆地区已有的液化调查数据为基础,基于逻辑回归方法建立了砂土液化概率判别公式,并给出了给定液化概率的临界标准贯入击数计算公式。通过回判和预判分析,并对比我国建规法的判别结果,对本文液化概率公式的可靠性进行了检验。4.采用台湾集集大地震液化数据进行检验分析。介绍了台湾集集大地震的背景资料,收集整理了我国台湾集集大地震液化调查数据,并对液化数据的分布特征进行了系统分析。利用台湾集集大地震液化数据作为应用实例,通过回判和预判分析,对本文液化概率公式的适用性和判别结果的可靠性进行了检验。结果表明,这一方法对台湾集集大地震液化数据的总体判别结果略显保守,但总体判别正确率较高,能够较好的反映实际场地液化情况。
[Abstract]:Sand liquefaction is a typical earthquake disaster phenomenon. The probability expression of liquefaction discrimination is one of the hot issues in geotechnical seismic engineering. The evaluation of engineering site condition is of great theoretical significance in engineering application. This paper is based on the existing liquefaction investigation data in mainland China. By using the method of logical regression analysis, the discriminant formula of sand liquefaction probability and the calculation formula of critical standard penetration number of given liquefaction probability are established. In addition, the discriminant results of the current code for seismic design of buildings in China are compared. The applicability and reliability of the liquefaction probability discriminant formula given in this paper are tested. The main work and research results are as follows: 1.The present research situation of sand liquefaction at home and abroad is summarized, and the earthquake damage phenomenon of sand liquefaction is summarized. This paper introduces the liquefaction mechanism and influencing factors of sandy soil, summarizes and comments on the research status of sand liquefaction discrimination methods at home and abroad, and briefly reviews the advantages and disadvantages and applicability of each discriminant method. The problems and shortcomings of the existing liquefaction discrimination methods for sandy soil are pointed out. The data of liquefaction investigation in the mainland of China are collected and analyzed, and the background data of 8 large earthquakes in the mainland of China are introduced. The existing liquefaction investigation data in continental area are collected and sorted, and the intensity of the existing liquefaction data, the standard penetration number, the distribution characteristics of sand depth and groundwater level are systematically analyzed. The basic data for the establishment of liquefaction probability discriminant method are provided. 3. The discriminant formula of sand liquefaction probability is established by using logical regression. The basic principle and application field of logical regression model are introduced. Four parameters of sand bed burial depth and groundwater level are taken as indicators of liquefaction evaluation. Based on the existing liquefaction survey data in mainland China, the probability discriminant formula of sand liquefaction is established based on logical regression method. The calculation formula of critical standard penetration attack number for given liquefaction probability is given. Through the analysis of back judgment and pre-judgment, the discriminant results of our country's construction planning method are compared. The reliability of the formula of liquefaction probability is tested in this paper. Finally, the liquefaction data of Jiji earthquake in Taiwan are tested and analyzed. The background data of Jiji earthquake in Taiwan are introduced, and the survey data of liquefaction of Jiji earthquake in Taiwan are collected and sorted out. The distribution characteristics of liquefaction data are systematically analyzed. Using the liquefaction data of Jiji earthquake in Taiwan as an application example, through the analysis of back judgment and prediction, The applicability of the liquefaction probability formula in this paper and the reliability of the discriminant result are tested. The results show that the overall discriminant result of the liquefaction data of Jiji earthquake in Taiwan is slightly conservative, but the overall correct rate is higher. Can better reflect the actual site liquefaction situation.
【学位授予单位】：中国地震局工程力学研究所
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU441;TU435

【参考文献】