| dc.description.abstract | The study is based on an actual soil profile from the Engineering Geological Database for TSMIP (EGDT), and then sets the geological conditions of the empirical model, mainly focusing on the soil liquefaction assessment of the sandy soil layer, referring to Andrus and Stokoe [1] made the following recommendations, or constraints: (1) groundwater table is situated at 0.5-6 m, (2) the soil’s Vs is within 100-200 m/s, (3) uncemented soils of Holocene (< 10,000 years), and (4) the middle point of the target (sandy) layer is no deeper than 10 m. In addition, the database in 17 N-Vs relationships was analyzed under the conditions of different earthquakes and different fines content, with a total of 1683 data points.
Mainly based on three soil liquefaction assessment procedures [1-3], 17 empirical relationships between shear wave velocity and SPT-N values [12-28], calculate each factor of safety, effectively observe the difference, and find out which is more conservative in engineering. Next, we developed the liquefaction assessment model which is based on the ratio of SPT- to Vs-based liquefaction factors of safety.
It shows that the probability distribution of R can be well modeled by the lognormal distribution with mean value = 1.76 and standard deviation (SD) = 0.59. This probability model can help us estimate soil liquefaction probability more easily, via an unsophisticated, deterministic calculation. Last but not least, the model is statistically robust according to the results of 105 verification tests. | en_US |