M.Sc. Tezi Görüntüleme

      In this study, the empirical equations were proposed for prediction of strength parameters as a function of index properties of regolithic soils in Trabzon City. The dominant lithology in the studied area is Eocene-Neojene aged basaltic, andesitic volcanic rocks. Plio-Quaternary aged altered regolith horizon overlying the volkanic rocks contains the unconsolidated mantle of weathered volcanic rocks (saprolite) and soil material. Thickness of regoliths including both of the soil itself and saprolite is about 5-10 meter in the region. In order to establish some relations between strength parameters and index properties, 45 soil samples were collected in Plio-Quaternary aged regolith in Trabzon city. In the study index properties, atterberg limits and grain size distrubition of clayey soils were determined in accordance with ASTM (1991). Coarse sieve and fine sieve methods were used for grain size analysis. These analysis shows that samples are composed of 1-3 % gravel, 11-20 % sand, 20-30 % silt and 35-50 % clay size materials. Specific gravity of the samples was determined between 2.36-2.61. Liquid limits are between 33 % and 90 %. The plasticity index extended from 10 % to 40 %. The value of activity was recorded as between 0.50 and 1.30. Clay materials were classified according to their plasticity index and liquid limit values and according to these properties the most of samples contain illite and kaolinite typed clay minerals. Result of the unconsolidated-undrained direct shear, cohesion of the soils was determined as between 10 and 51 kPa. Thus, the tests show those friction angles are between 7 and 40. To determine the statistical relations between strength parameters and physical properties of clayey soils, multiple regression analysis was carried out. According to the multiple regression analysis results, the negative correlations characterized by the regression coefficient of 0.80 were determined between friction angle and PI/LL. However, the positive logarithmic relations between cohesion and PI/LL are determined. This relation is characterized by a regression coefficient of between 0.80.