In this study, a low-cost, rapid and qualitative evaluation procedure using a fuzzy set analysis for assessment and prediction of liquefaction potential of saturated sandy grounds is presented. Eight items affecting liquefaction resistance of ground including geology, geomorphology, seismic (relative site amplification and intensity increments), and geotechnical items (sandy layers thickness, water table level, thickness of surface layers, and type of soils) are considered to express the basic characteristics of liquefaction potential of the ground. These items are chosen and established from a review of the various literatures, engineering judgment, available statistical data, and previous observations of liquefaction in the world. A set of evaluation criteria was established or selected for each item and a total of eight factors is used in the proposed evaluation system. In the proposed evaluation system, liquefaction potential of ground is assessed and expressed in linguistic terms based on the considered criteria. Then the linguistic data is analysed by using fuzzy sets. The liquefaction index is defined for assessment of liquefaction potential of soils.
An example of application of the method is presented to liquefaction potential analysis of saturated soft ground in northwestern Iran (Gilan plain) . The studied area was suffered catastrophic earthquake in June 1990, and the properties of the earthquake were widespread liquefaction, several huge landslides and more than tens other slope failures.
This study revealed that, the proposed method is able to predict the liquefaction potential of the ground for preparation of hazard potential maps and zoning, which is useful for general hazard assessment and delineation of areas.

View full abstract

N value data stored in a geotechnical database can be used for the analysis of subsurface structures from an engineering geological point of view. The purpose of this article is to reveal the subsurface geotechnical properties of two areas based on the N value data, namely, the Tokyo coast area in central Japan, and the Kumamoto plain in southwest Japan. The geotechnical databases for Tokyo and Kumamoto areas consist of 164, 870 borehole logs, respectively. N value data of these areas were analyzed using the clustering and the fuzzy contouring. The clustering of borehole logs based on the N values from surface to 30 meters depth with an interval of 1.0 meter revealed that, in both areas, borehole logs with similar vertical changes of N values tend to be distributed close. The results of the clustering showed that seismic engineering properties including empirical equation between the N values and the S wave velocity, predominant frequency, and amplification characteristics for ground shake differ from each of the clusters.
Fuzzy contouring, which had been proposed by the authors, is based on a polynomial expression with fuzzy coefficients. The method was developed so that the estimated width is affected by the distance of sample data from the interpolated point, and the deviation of the sample data's distribution. The irregularly distributed N values at the same depth from the ground surface were interpolated using the fuzzy contouring for the Kumamoto plain. The interpolated area could be classified based on the estimated width of N values. It was concluded that the clustering and the fuzzy contouring of N values are considered useful methods to identify the area where the geotechnical properties change.
Furthermore, the N values were used to determine the seismic engineering basement in the shallow subsurface and to estimate the S wave velocities above the basement. Exploiting the multi-reflection theory and the estimated S wave velocities, seismic intensity at each boring site which might be induced by an earthquake with assumed magnitude and epicenter distance was predicted.

View full abstract