Collapsibility study for tropical soil profile using oedometric tests with controlled suction

Abstract

The aim of this research is to present and discuss the results of oedometric tests under controlled matric suction, for lateritic soil samples from an unsaturated soil profile, typical of 14% of the region of Campinas, in the State of Sao Paulo, Brazil. The soil profile consists of colluvial, lateritic and collapsible silty clay, with approximately 6.5m thickness. The water level is approximately 17m. The textural, structural and compositional characteristics of the soil profile demonstrate two distinct sources in the upper layer (from 1.5m to 5.5m) as well as in the lower layer (at 6.5m). The former has a sedimentary source and the later has an igneous and/or metamorphic source. However, in the lower layer, the structural characteristics of the rock source prevail over the upper layers, with the evident presence of igneous origin minerals. The hematon/goetan + argilon aggregates are the most common, followed by ferruginous concretions, magnetite grains and pyroxene pseudomorphs, suggesting a source near the Diabase. All these components are involved in a matrix with clay minerals, oxides and iron hydroxide and an expressive quantity of voids. Although there is the presence of cimentation, from iron oxides, the particles, surrounded or not by hematite and goethite, and microagregates are linked by capilarity forces. This fact indicates that the soil suffers disaggregation when wet. Disturbed and undisturbed samples were collected from this profile, at a 1.5m depth. Six specimens were made from the undisturbed samples and submitted to oedometric tests. The specimens were wet and submitted to axis-translation technique, with pressure equipment and a pressure control panel. The imposed matric suction values were: zero, 25kPa, 50kPa, 100kPa, 200kPa and 400kPa. The levels of net stress applied were 6kPa, 12.5kPa, 25kPa, 50kPa, 100kPa, 200kPa, 400kPa, 800kPa and 1600kPa. Experimental results, obtained by oedometric tests, indicate collapse index varying between 2% and 18.7%; indicating collapsibility. Moreover, the experimental results were inserted into a theoretical model to estimate the LC curve and evaluate its application on this lateritic soil. intensity, allowing the relative displacement between particles, so that they start to fill the soil structure voids (Dudley, 1970). Collapse study in tropical soils is usually carried out using single or double conventional oedometer tests, without taking into account the influence of matric suction on soil behavior. The use of oedometer tests with suction control is vitally important for collapse study in unsaturated soils. By using this technique, it is possible to predict soil behavior for known values of matric suction, eliminating the problem of not knowing the initial moisture content of the specimen, which is a value that significantly influences the increase of specimen collapse. Conventional oedometer tests performed on studied soil showed that it is a collapsible soil when tested using field moisture. The study of thin and impregnated soil plates at various depths showed that the link between grains of soil is basically due to the presence of matric or adhesiveness without the presence of cementations that may help stabilize the soil structure. Thus, due to the great importance of matric suction on the mechanical behavior of the studied material, an option was made to use oedometer testing with suction control, making it possible to verify the real contribution of this variable in the resistance of the material. Therefore, this paper analyzes the results of oedometer tests carried out with suction control for the superficial layer of the colluvial and lateritic soil profile at the Experimental Field of Soil Mechanics and Foundations, UNICAMP EFSMF, located in Campinas / SP, Brazil, and aims to assess the influence of matric suction in the hydro-mechanical behavior of this soil and to verify the applicability of the elasto-plastic model proposed by Alonso et al (1990) for the samples tested. 2 MATERIALS AND METHODS