基于振动台试验开展砂土液化的机理与判别研究

发布时间：2018-08-07 15:04

【摘要】：砂土液化是岩土工程的重要研究课题,由动荷载引起的砂土液化造成地基失效、地面喷砂冒水、建筑物破坏等现象,对国民经济造成了巨大损失,使人民生命财产受到严重威胁。随着人们对于液化问题的不断深入关注,在实际工程中发现了大量非饱和砂土场地的液化现象,然而目前对于非饱和砂土液化的研究还相对较少,因此研究非饱和砂土的液化机理对于工程应用至关重要。本文基于振动台试验开展了饱和度分别为100%,95%,90%,85%,80%的五组砂土模型地基的液化试验,通过试验结果分析发现:饱和度是影响砂土液化的重要因素之一,随着饱和度的降低,砂土地基的液化抗性随之增大;砂土地基的液化顺序是自下而上进行的,即随着循环荷载的持续进行,砂土地基中的孔隙水不断地从下往上涌动,中下层的砂土首先液化,上层砂土随后液化,直至顶层砂土发生喷砂冒水现象,整个砂土模型地基全部达到液化;各饱和度的砂土模型地基在振动结束之后,超静孔隙水压力均会发生一定程度的消散,超静孔压消散的速度随着埋深的增大而加快。其次,由于目前对于液化判别尚未形成统一认识,各判别方法均存在一定的局限性,不具有广泛适用性,判别准确率不高。本文基于Bayes判别分析理论,根据唐山大地震和广东三水地震的40组砂土液化样本构建了数据库。首先,以震级M、地面加速度最大值gmax、标准贯入锤击数N63.5、比贯入阻力Ps、相对密实度Dr、平均粒径D50和地下水位深度dw七个评价指标建立了七因素Bayes判别分析模型。继而,依据七因素Bayes模型中各评价指标的影响权重,提出了以震级M、地面加速度最大值gmax、比贯入阻力Ps、相对密实度Dr和平均粒径D50五个因素为评价指标的五因素Bayes判别分析模型和五因素Fisher判别分析模型。最后,通过系统地对比五因素模型和七因素模型的砂土液化判别结果,发现五因素模型的判别准确率均高于七因素模型,即本文所提出的五因素模型用更少的评价指标可达到更高的液化判别准确率,既避免了次要因素影响判别函数的稳定性,又节约了建设成本,更有利于实际工程中的液化判别。
[Abstract]:Sand liquefaction is an important research subject in geotechnical engineering. The liquefaction of sand caused by dynamic load causes ground failure, sand blasting and water blasting on the ground, building destruction and so on, which has caused huge losses to the national economy. Put people's lives and property under serious threat. As people pay more and more attention to liquefaction problems, a large number of liquefaction phenomena of unsaturated sand sites have been found in practical engineering. However, the research on liquefaction of unsaturated sandy soils is relatively rare. Therefore, it is very important to study the liquefaction mechanism of unsaturated sand. Based on the shaking table test, the liquefaction tests of five groups of sand model foundations with saturation of 100% and 85% and 80% have been carried out in this paper. Through the analysis of the test results, it is found that saturation is one of the important factors affecting the liquefaction of sand soil, and with the decrease of saturation, The liquefaction order of sand foundation is from bottom to top, that is, with the continuous cyclic load, the pore water in sand foundation continuously flows from bottom to top, and the sand in middle and lower layers is liquefied first. The upper sandy soil then liquefies, until the top sandy soil occurs sand blasting, and the whole sand model foundation reaches liquefaction, and the excess pore water pressure of each saturation sand model foundation dissipates to a certain extent after the vibration is over. The velocity of excess pore pressure dissipates with the increase of buried depth. Secondly, because there is no unified understanding of liquefaction discrimination at present, each discriminant method has some limitations, it is not widely applicable, and the accuracy of discrimination is not high. Based on the theory of Bayes discriminant analysis, the database of 40 groups of sand liquefaction samples from Tangshan earthquake and Sanshui earthquake in Guangdong province was constructed. Firstly, a seven-factor Bayes discriminant analysis model is established based on seven evaluation indexes: magnitude M, maximum ground acceleration g _ max, standard penetration hammer number N _ 63.5, specific penetration resistance P _ s, relative density D _ (50), mean particle size D _ (50) and depth of groundwater level dw. Then, according to the influence weight of each evaluation index in the seven-factor Bayes model, A five-factor Bayes discriminant analysis model and a five-factor Fisher discriminant analysis model are proposed, in which magnitude M, maximum ground acceleration g _ max, specific penetration resistance P _ s, relative density Dr and mean particle size D _ 50 are used as evaluation indexes. Finally, by comparing the results of sand liquefaction discrimination between the five-factor model and the seven-factor model, it is found that the accuracy of the five-factor model is higher than that of the seven-factor model. That is to say, the five-factor model proposed in this paper can achieve higher liquefaction discrimination accuracy with fewer evaluation indexes, which can not only avoid the influence of secondary factors on the stability of discriminant function, but also save the construction cost, which is more favorable to liquefaction discrimination in practical engineering.
【学位授予单位】：浙江海洋大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU441;TU435

【相似文献】