Análisis del comportamiento tenso-deformacional de arcillas terciarias sobreconsolidadas en el entorno del Campo de Gibraltar

Abstract

The term ‘Flysch’ encompasses a set of sedimentary deposits made up of rhythmic alternations of rocky materials (sandstone, marl-limestone, etc.) with less competent clayey materials. These materials are present in many parts of the planet, always associated with orogenic processes. From a geotechnical point of view, Flysch-type materials are highly difficult to sample. In general, even the clayey materials of the Flysch tend to present intermediate behaviours between soft rocks and very firm soils, which gives them enough resistance to complicate their sampling using conventional samplers. For this reason, it is necessary to resort to more complex techniques to obtain samples on this type of land, and this is one of the reasons why these materials are generally little studied. These materials are present in many terrestrial and maritime infrastructures, which are frequently conditioned by the special nature of these materials (Manzano et al. 2020). Landslides, problems associated with the stability of the foundations or stability problems in the excavation of tunnels are some of the geotechnical problems that are frequently detected in this type of material. In this thesis, a detailed physical, chemical, and mineralogical characterisation of the clayey materials of the Campo de Gibraltar Complex (Southern Spain) were carried out. From the grading point of view, these materials were shown to possess a high percentage of fines (69.92 % on average), although most of them had a certain coarse content, due to the difficulty in dispersing the clayey aggregates. Regarding plasticity, the samples analysed showed a mean value of the liquid limit of 38.2 with a mean plasticity index of 17.5. Regarding the natural moisture content, the samples analysed presented mean values below the plastic limit (mean moisture = 15.7%). This means that the analysed soils were classified mostly as inorganic clays of low plasticity (Manzano et al., 2022). In addition, X-ray diffraction tests have also been carried out to establish the mineralogy in the samples. In these tests, quartz was the main mineral present in all cases, followed by a significant percentage of illite. Clinochlore also appeared widely as a major mineral, while both calcite and dolomite appeared to a lesser extent. Minerals such as beidellite, palygorskite, fluorapatite, chamosite and nontronite appeared as accessory mineralogy of this type of clay. From a strictly chemical point of view, the samples analysed presented low sulphate and soluble salt contents, and hardly any organic matter. Finally, the average carbonate content was 5%, with significant differences between samples. In addition to identifying and studying the composition of the soil, tests were carried out to determine the resistant properties of this type of soil. To this end, triaxial tests (CU and CD types) were carried out in addition to oedometric tests. From these tests, the characteristic values for each of the samples were extracted so that they could be represented according to three different constitutive models: Mohr-Coulomb Model, Modified Cam Clay Model and Hardening Soil model. From the tests, the samples were verified to show a hyperbolic behavior under triaxial conditions with a progressive hardening as the deformation increases. In addition, models were made using finite elements of the triaxial tests to check the goodness of fit for each of the established models. In all the modelling carried out, it was verified that the Hardening Soil model is the one that adjusts most precisely to the real test data, this adjustment being significantly better for low unit strains. The Cam Clay and Mohr-Coulomb models did not show reasonable fits other than in the first part of the elastic behaviour. In the final part of the thesis, an analysis was carried out using finite elements of a real test of axial static load on a micropile. The results obtained in this test were compared with the results obtained for this model, observing how the best fit is obtained when the terrain is modelled with a Hardening soil type constitutive model, with the differences between the model and the real test data being (around 3%).