The vast territorial expanse of Brazil imposes the need for an equally extensive road network. A crucial aspect in designing and sizing pavements is related to regional factors, which encompass specific conditions of geology, climate, and terrain. These elements result in weathering processes capable of generating soils with distinct characteristics. Despite differences in weathering processes in the northern regions, with a humid tropical climate, and in the central region of Brazil, with a semi-humid tropical climate, both exhibit tropical soils, some with lateritic behavior and others not, due to the occurrence of various processes of formation and alteration of these soils. Soils, influenced by variables such as mineralogy, particle size distribution, and the degree of associated weathering, exhibit diverse mechanical behavior. On the other hand, crushed materials, with a relatively stable mineralogical composition, are strongly influenced by particle size distribution and the shape index. These considerations raise questions about the most appropriate approaches for flexible road pavement structures, as these pavements are intrinsically linked to stiffness interactions between layers with different properties. For a better understanding of the aspects associated with the soils of the Federal District (DF) and the state of Roraima (RR), and graded aggregates of the DF, this research proceeded with the analysis of different materials, aiming at micro and macrostructural characterization, as well as the mechanistic behavior of these materials. It was noted through the results that classifications and characterizations provided information that explains the different mechanical behaviors of soils and graded aggregates. Thus, not only the state of stress, but also the origin and formation of materials affect deformability behavior. In this way, the importance of a decentralized view of purely mechanical tests was highlighted, also seeking information on the genesis of materials.

The interaction between soil dynamics and the atmosphere represents a significant challenge and an area of increasing interest in geotechnical engineering. This dissertation focuses on modeling the behavior of unsaturated slopes subjected to intense precipitation events, exploring the applicability of analytical models. The theoretical foundation includes Darcy's Law and the Richards Equation, supplemented by recent advances in modeling flows in unsaturated soils, as exemplified by the approach of Cavalcante & Zornberg (2017). The study assesses the accuracy of these models in predicting the transient behavior of unsaturated slopes, validating the predictions with experimental data obtained from a scaled model of a slope. This process involves understanding how variations in the internal energy of fluids, hydraulic conductivity, and the soil-water retention curve influence water movement and, consequently, the stability of slopes. The dissertation explores the application of a probabilistic approach to assess the stability of unsaturated slopes, employing the Rosenblueth Point Estimate Method (PEM) implemented in the Wolfram Mathematica software. The analysis focuses on determining the transient failure probability and the reliability index, incorporating random variables such as dry unit weight, hydraulic conductivity, the hydraulic adjustment parameter, and the soil's effective friction angle. Sixteen simulations were conducted, considering the random variables and their distributions to estimate the safety factor under various precipitation and humidity conditions, calculating the failure probability for different tests and depths. This methodology allowed for identifying critical moments of increased failure probability, emphasizing the importance of initial moisture conditions and the effects of preceding rainfall.

Unpaved roads are important roads that often connect urban to rural areas. Geosynthetics are reinforcing materials that are used in these roads in order to generate, for example, a better structure support capacity, less deformations, better distribution and stress transfer. The improvement in mechanical performance makes performing surface maintenance less necessary and thus the costs related to repairs decrease in the long term. When it comes to reinforcement, the most applied types of Geosynthetics are geogrids and geotextiles. Numerical analyses to predict certain behaviours of the structure are efficient and capable of studying previously the benefit in the use of this reinforcing material. This research uses numerical to simulate an unpaved road under cyclic loading. The structure is formed by a fill layer, a compressible subgrade and a geosynthetic layer placed at the fill-subgrade interface. Surface maintenance simulations were also performed and analyzed. The methodology used was based on the Stiffness Degradation Method and the software used is based on the finite element method. Comparisons between numerical and experimental results were made and it was concluded that most of the simulations showed a good agreement with the experimental test result. The numerical results of surface displacements were the ones that best agreed with the experimental ones in contrast with the vertical stresses in the subgrade probably due to the high level of plastification inside the structure along the tests.

Geotechnical engineering has utilized technological advancements and numerical analyses to minimize errors in soil property measurements and achieve more accurate results. Methods from materials science have contributed to the analysis of soil properties at the grain scale, enabling morphological characterizations such as X-ray micro-computed tomography (µCT). In this context, this dissertation aims to implement algorithms and develop scripts in Python to process two-dimensional images generated by X-ray micro-computed tomography. The samples analyzed included sand, iron ore tailings, and a composite of iron ore tailings with polymer, aiming to analyze mesoscale morphological parameters such as sphericity, granulometry, porosity, and void ratio. The research highlighted that the configuration of the images in the micro-CT scanner and the pre-processing of the images can significantly influence the results. The use of Python for analyzing the morphological parameters of geotechnical samples proved to be relatively feasible, provided that the specificities of the micro-CT images of the studied sample are considered. 

Load testing is the primary means of mitigating uncertainties in deep foundation geotechnical projects. In this regard, dynamic load testing (DLT) is a highly attractive solution as it allows for rapid results with lower financial costs compared to static load testing. DLT is interpreted through CAPWAP analyses, which utilize signal matching techniques between field readings and those obtained from a pile-soil model. This technique involves trial and error, and interpreting the results requires good judgment from the engineer conducting the analyses and inputting the soil and pile parameters into the model. The most important parameters in the analysis are the shaft quake (qs) and damping factor (Js), which are commonly assumed to be constant along the pile shaft, regardless of soil stratigraphy (variation in soil type and mechanical properties). In other words, CAPWAP analyses in this approach become highly dependent on experience and subjectivity.
The aim was to establish correlations between the undrained shear strength (Su) and the qs and Js of the soil along the shaft of continuous flight auger (CFA) piles. It also aimed to correlate qs and Js with the penetration resistance index (N). To achieve this, Vane Shear Tests (VST) and Standard Penetration Tests (SPT) were conducted, along with a static load test and two dynamic load tests on piles at a construction site in Sinop, Mato Grosso, Brazil. The soil in the city has an alluvial origin and consists of poorly consolidated sediments. The terrain at the site comprises cohesive soil up to a depth of 15 meters (inorganic clay), transitioning to cohesionless soil (silty gravel, sandy silt and sand) down to a depth of 32 meters (limit of the boreholes). The groundwater table was observed at a depth of 2 meters.
The results of the tests were analyzed, and some correlations between soil properties and the qs and Js parameters obtained from the CAPWAP analyses were observed. Within the limitations of the analyses conducted, it was concluded that the CAPWAP model parameters are indeed not constant throughout the depth and vary according to the type and mechanical properties of the soil. It was found that most of the results diverged from those expected in the literature, possibly due to the different soil properties in Sinop, Mato Grosso, and the type of pile analyzed (CFA).

The mineral industry faces the challenge of ensuring sustainable mining with engineering practices that can reduce the potential damage associated with the operation of geotechnical structures. Based on that, the construction of experimental landfills is considered as an alternative to understand the performance of filtered tailings during compaction. This study addresses the construction of an experimental landfill of filtered flotation iron ore tailings constructed with the purpose of evaluating the behavior of the structure in terms of geotechnical properties and  the material workability for future construction of a filtered tailings pile. The layers were executed in order to present a minimum compaction degree (GC) of 98% and tests were carried out for quality control in each layer. In addition to piezometers installation, geophysical campaign, field tests and collection of disturbed and undisturbed samples were performed for the laboratory tests. The required compaction degree required was achieved with five turns of the road roller and there was no issues related to the operation. There was a decrease in water content during the filtering process resulting in some layers compacted below the optimum moisture content. Tailings was characterized as a fine silty sand, poorly graded and with no plasticity. The saturated hydraulic conductivity obtained in the triaxial tests is 10-3 cm/s and the determination of the retention curve using the filter paper method was adjusted considering the results of the reconstituted samples with a compaction degree of 95% and in situ samples (GC > 98% ). The CID and CIU triaxial tests in the in situ samples demonstrated high shear strengh (friction angle of 37º), a dilatative behavior and without excess poropressure generation. The CIU triaxial tests performed in samples with compaction degree of 95% indicate the generation excess poropressure, therefore this degree of compaction is not suitable for future construction of the tailings dump. The CPTu tests confirm the high shear strength of this filtered tailings, the dilatant behavior and no liquefaction susceptibility. The geotechnical instrumentation demonstrates that the positive pore pressure generation during the landfill construction was very low due to the moisture content below the optimun moisture during compaction.

The increase in greenhouse gas emissions, ocean pollution, large-scale deforestation, increased energy, and hydraulic demand, as well as other excesses produced by human action, contributed to the unnatural acceleration of climate change. As a consequence of climate change, it was observed that there has been a significant increase in the periodicity and intensity of climate phenomena, including mass movements induced by intense precipitation events. These phenomena, naturally, are processes of the genesis of the Earth's surface. However, they present a severe geological risk due to the magnitude of destruction caused in urban environments, where there is a civil occupation in hillside regions. In this context, several methods based on computational tools were developed to help understand the boundary conditions of the water flow on the slope resulting from rain, which ultimately determines its behavior and mechanical performance. However, physical modeling also showed promise due to the possibility of reproducing failure mechanisms in the laboratory, facilitating the perception of the water flow process, visualizing changes and effects generated by humidity, and monitoring rupture processes. Based on the range of possibilities for analyzing flow and failure phenomena that can be produced in a physical model, the objective of this Doctoral Thesis was to carry out an experimental analysis through the construction of a physical model of a 1g small-scale slope for the evaluation of failure indicator parameters during the simulation of superficial landslides, induced by artificial precipitation on a sandy slope under unsaturated conditions. For this, Experimental Equipment with the implementation of an artificial rain system was built to enable the physical modeling of the slope with the application of a monitoring system to monitor the hydraulic parameters during the experimental program. As a result, it was possible to identify fundamental principles for studying the infiltration phenomenon and behavior patterns even in different simulated scenarios.

In recent years, records of human losses associated with mass movements have increased significantly, becoming a latent hazard to communities in mountainous regions. Bearing in mind that the rupture of a massif is not always achieved under saturated conditions and that in some cases, the analysis of the unsaturated phenomenon may be more reliable, this research proposes an integrated platform for assessing the probability of failure in the unsaturated condition, which is the main input for assessing the risk of landslides triggered by precipitation events. In this work, the uncertainties associated with the physical, hydraulic and resistance parameters that define slope stability are incorporated by applying the First-Order Second-Moment Method (FOSM). To do this, the analytical stability model proposed by Cavalcante and Mascarenhas in 2021 is implemented, which was developed for unsaturated porous media considering a transient analysis. To implement the model, a computational routine was developed in the Wolfram Mathematica programming language. The methodology proposed in this research is validated using information from Gerscovich's work published in 1994, which analyzed a landslide associated with precipitation in the Rio de Janeiro region. The implementation of the methodology developed in this thesis demonstrated the importance of migrating from deterministic analysis to probabilistic analysis, making it easier for engineers to use tools that allow them to determine the statistical moments of the properties involved in the physical phenomena and their importance in the scenario chosen for stability analysis. The use of FOSM produced good results in obtaining the probability of failure in the cross-section of the slope, and had a positive performance in terms of computational cost. The analyses carried out for the case study showed the importance of analyzing the previous rainfall to define the initial state prior to the failure event, as this scenario is indispensable for obtaining more realistic hazard profiles. Finally, it is shown that the transient probabilistic approach to slope stability is an excellent alternative for assessing the hazard associated with the occurrence of a mass movement, since, depending on precipitation events, the state of humidity and stress varies within the massif, also changing the statistical dispersion of physical properties along the slope.

Railway ballast modeling can be performed by different approaches, through continuous or discrete models, which have their comparative advantages and disadvantages. Continuum mechanics-based methods often rely on numerous empirical parameters and require significant laboratory effort for calibration. Conventional discrete element methods have a high capacity to represent railway ballast; however, due to physical inconsistencies in their formulation, they may encounter similar issues. This study aims to adapt and propose the Hybrid Lattice-Discrete Element Method (Hybrid LED) for virtual modeling of railway ballast tests. The advantages of employing this technique include: (i) simplified determination of input parameters and optimization of laboratory efforts, (ii) realistic representation of complex geometries of railway ballast particles, (iii) visualization of micromechanical phenomena, such as particle fracturing, and (iv) monitoring of state variables throughout simulations, such as stress state, breakage indices, porosity, and particle size distribution. Initially, material parameters were defined based on laboratory test results obtained from the literature. Then, particle generation, Voronoi discretization and packing algorithms were used to build models of railway ballast samples. These models were used to simulate mechanical tests, namely single particle compression, confined uniaxial compression, monotonic triaxial compression and cyclic triaxial compression. There was consistency between the results and the empirical observations reported in the literature. In addition, variations in particle size distribution were observed during the simulations, and the causes of failure, primarily attributed to particle slippage or breakage, were investigated. Moreover, predominant particle breakage modes were assessed, such as corner breakage or crushing. By analyzing these elements together, a detailed understanding of the mechanical behavior of the studied material may be obtained. The primary drawback of the method was its computational cost, which, although not prohibitive, is higher than conventional methods. However, this challenge can be overcome through the utilization of computational resources such as clusters and cloud computing. In conclusion, the Hybrid LED method proves to be effective for simulating laboratory tests on railway ballast, capable of representing complex geometries and various boundary conditions.

In response to the challenges presented by the continuous increase in Construction and Demolition Waste (CDW) generation in expanding urban contexts, this study proposes an innovative technical approach for the reuse of these materials. The objective is to explore the potential of CDW as alternative aggregates in asphalt mixtures, an area that has been gaining increasing attention in the field of civil engineering.
The research focuses on two specific granulometries of CDW. This choice allowed for a detailed comparative analysis of how different aggregate sizes can influence the behavior and properties of asphalt mixtures. The morphological characterization of CDW, performed using 3D scanning technology and Scanning Electron Microscopy (SEM), was a key aspect of the study, providing valuable insights into the texture and shape of the waste. The SEM analysis, complemented by Energy Dispersive Spectroscopy (EDS), enabled a detailed observation of the CDW surfaces with and without the additive, revealing the presence of adhered materials such as clays, fissures caused by adhered mortar, and fractures of the material during crushing. Additionally, X-Ray Diffraction (XRD) tests were conducted to determine the mineralogical composition of the CDW.
An innovative aspect of the research was the application of the TERRAFIX additive to the CDW through an absorption process. This technique aimed to improve crucial characteristics of the CDW, such as waterproofing and adhesion, making them more compatible for use in asphalt mixtures. The treatment with TERRAFIX represents an important step in the preparation of CDW, as it modifies their surface properties and can enhance interaction with the asphalt binder.
The laboratory tests focused on evaluating the mechanical performance of asphalt mixtures incorporating the treated CDW. Various tests were conducted to determine parameters such as stability, wear resistance, moisture damage, fatigue, resilient modulus, and tensile strength of the mixtures, among others. The results indicated that asphalt mixtures with treated CDW showed promising performance, suggesting that this approach can offer a viable alternative to conventional aggregates, both technically and economically.
This study can significantly contribute to the field of civil engineering, providing a solid foundation for future research and practical applications. The reuse of CDW in asphalt mixtures not only presents a technical solution for managing this waste but can also lead to significant advances in terms of resource efficiency and innovation in construction materials.

This work aims to contribute to the knowledge of self-drilling tubular micropiles
installed in the soil of Brasilia. This pile is one of the “Cast-in-place” type and for its
molding a rotary drilling with simultaneous injection is used. Eight load tests were
carried out on isolated piles, varying the parameters of injection pressure, diameter and
length; and four other load tests in groups of two, three, four and five piles, all with the
same executive characteristics. The diameters analyzed for the isolated piles were 200
and 260 mm, and the diameter used for the group piles was 200 mm. The injection
pressures used were 200 kPa and 250 kPa for isolated piles and 200 kPa for group piles.
The length used for the isolated piles was 8 m and 12 m and for the group piles the
length was 8 m. The efficiency of the group of piles at 10% and 5% of settlement
normalized by diameter was also analyzed. It can be concluded from this work that the
adequate choice of the executive characteristics of the pile allows the optimization of
the foundation system. This type of pile proved to be a viable alternative for porous
soils with low resistance to penetration (Nspt) such as those in Brasilia, as it can offer
technical, economic and environmental advantages in relation to other types of piles,
mainly those related to their high execution speed and low cost.

The computational program MeDiNa allows the verification and empirical mechanistic design of asphalt pavement structures through the application of the theory of multiple elastic layers. Among its functionalities is obtaining parameters of deformability of layer materials, which allow evaluating the behavior of existing pavements by means of retroanalysis of deflectometric basins. Therefore, the objective of the present study was to initially evaluate the consistency of the resilience modules obtained by the program, as well as the adherence of the simulation of the pavement service life to the conditions observed in the field. For that, a set of data collected in 8 (eight) Sampling Units (homogeneous segments with 1 km in length) different from the DNIT, located in the region close to the Federal District, including the characteristic traffic, the Longitudinal Irregularity Index (IRI), the percentage of cracked area of the pavement, thickness information and physical characterization of its layers obtained through inspection wells, as well as data from deflectometric basins measured by FWD equipment. After verifying the hypothesis that the Transfer Function of the program offers conditions of use for the region, a second stage was established. This stage consists of presenting guidelines, based on precision levels, for the best use of MeDiNa in asphalt paving projects, as well as its use for structural quality control of these pavements, since the tool has a useful life forecast model based on in criteria of rupture due to fatigue and permanent deformations, which reflect in cracking in the asphalt coating and sinking of the wheel track.

After the accidents at Fundão Dam and Dam 1 of the Córrego do Feijao Mine, Law No. 14,066, of September 30, 2020, decreed the need to decharacterize all upstream tailings dams in operation in the country until February 25, 2022. Since this is an activity that has been poorly studied and with incipient definitions of construction processes, mapping alternative executive methodologies for the decharacterization of tailings dams is of utmost importance. In this context, the aim is to evaluate the susceptibility to liquefaction, using the Olson Method, of a typical section of a tailings dam raised upstream, considering the total removal of material from the structure as the decharacterization methodology. The Olson Method is an empirical technique that uses correlations of SPT and CPT test results to analyze the liquefaction potential of soils. This method is divided into three steps: in the first, it aims to verify the susceptibility of liquefaction of the material, then, it aims to verify the possibility of triggering liquefaction and, finally, the Safety Factor of the liquefied material. Two different construction methodologies were evaluated: removing material from the dam in the downstream to upstream direction and in the opposite direction, from upstream to downstream. The results indicate slightly higher Safety Factors in the analyses considering material removal from downstream to upstream. The higher elevations present lower Factors of Safety in both scenarios, where there is a high relevance of the tailings strength ratio in the overall stability of the structure. As excavation occurs, the strength parameters of the starter dam and raising become predominant, increasing the Factors of Safety in all types of analysis and causing the results of the liquefied analyses to approach the peak scenario values. It can be seen, therefore, that with the progress of the decharacterization, there is no significant influence of the tailings strength properties on the overall stability of the structure, a result in line with what was expected and that corroborates with most of the decharacterization works currently being executed in the country.

The disorderly growth of the city population resulted in an urban policy that perpetuates social inequalities. Thus, public space has been abandoned, reinforcing the idea of private property as the only safe place, and increasing economic distances with irregular occupation of peripheral and environmentally fragile areas. The urban expansion process depends on the incorporation of new areas and changes in land use, so that the city has a spatial organization that meets the demands of the new population. The general objective of this thesis is to explore cartographic methods and models to aid sustainable urban planning, based on the structuring of a progressive geotechnical mapping procedure, using geotechnologies and the theoretical basis of Geotechnical Cartography. For this purpose, data on different approach scales were used, with the aim of seeking suitable cartographic products to assist decision-making aimed at urban suitability. This study presents a quantitative approach and, from the point of view of technical procedures and based on satellite image classification, using methodologies such as Progressive Detailing and AHP (Analytic Hierarchy Process). The results obtained aimed at characterizing the susceptibility to hydrogeological processes, mass movement and erosion that served as the basis for the construction of the Geotechnical Chart of Aptitude for Urbanization of the Pratagy River Basin, respecting the legally imposed environmental restrictions.

Geotechnics is an important field due to the large scale and significance of its structures for society, such as dams, deposit piles, road cutting and filling, foundation of civil structures, and many other works where geotechnical knowledge is necessary for their execution. However, serious impacts on society are associated with this type of work, where the risks of failure can generate significant financial, environmental, and human life losses. Advances in geotechnical research allow for the improvement of numerical methods and constitutive models, seeking a better representation of the behavior of soil exposed to different boundary and loading conditions. Thus, models that can be used in the evaluation of geotechnical problems considering the particularities observed in tropical soils have been presented. Following this, an applied discussion shows the application of the Shear Strength Reduction (SSR) method considering the Modified Cam-Clay Hardening (MCC) critical state model. An evaluation of the performance of critical state models in laboratory testing was also carried out, indicating limitations of the MCC model for assessing the behavior of tropical soils. Finally, a simplification of the Madus Sub-loading model (Cordão Neto et al., 2009) was implemented in a commercial software for the development of a constitutive model that considers the critical state concept and the effect of sub-loading.

Leakage through mechanical damages is a critical concern in geosynthetic systems, particularly with the use of PVC (polyvinyl chloride) and HDPE (high-density polyethylene) geomembranes. Mechanical damages can occur during installation, operation, or as a result of external factors causing punctures, cuts, or abrasions. This study aimed to evaluate water leakage through mechanical damages in smooth PVC and HDPE geomembranes with a thickness of 1 mm. Longitudinal cuts of 5 mm, 20 mm, and 50 mm were made using a blade, and holes of 2.2 mm and 5 mm were created to simulate damages in the geomembranes during installation and soil cover placement. Leakage tests were conducted using equipment developed at the Geotechnical Laboratory of the University of Brasília, applying hydrostatic pressures of 20 kPa, 100 kPa, 200 kPa, 400 kPa, and 800 kPa, equivalent to water heads ranging from 2 m to 80 m of water. The study has applications in geotechnical engineering projects such as dams, channels, and waste disposal areas, where leakage may occur under high pressures. The objective was to understand the factors that influence leakage rates and to develop methodologies to mitigate potential risks. The results showed that leakage intensity depends on various factors, including hydrostatic pressure, water temperature, and geometric characteristics of the damage (size and shape of the mechanical damage), which significantly affect the leakage rate. Larger damages tend to result in higher leakage compared to smaller ones, and hole damages appear to be more critical than linear cuts. The presence of sharp edges or irregularities in the damaged area can also influence leakage. PVC geomembrane barriers showed lower leakage compared with HDPE geomembranes. To mitigate leakage, it was observed that applying powdered bentonite under the geomembrane and in direct contact with the damage helped seal the leakage (using hydraulic barrier GM/GCL-D-F5). Therefore, understanding the behavior of PVC and HDPE geomembranes when subjected to mechanical damages is crucial for the design and performance of geosynthetic liner systems. Continued studies to minimize leakage rates through appropriate installation practices, material selection, and timely repairs are important to ensure the long-term effectiveness and reliability of geomembranes in various geotechnical applications

The implementation of waterproofing systems during the construction of dams, landfills, and artificial channels is crucial to avoid fluid infiltration into the foundation soil, which can lead to structural damage and other potential hazards. Geomembranes are widely used in civil engineering as they offer excellent waterproofing capabilities, low permeability, and favourable mechanical properties. Like other construction materials, evaluating the strength at the interface between geosynthetics and in-contact material is necessary to ensure enough resistance to potential failures. Various laboratory tests can be conducted to ensure adequate interface resistance, such as direct shear, ring shear, and inclined plane. These tests determine the friction angle, a critical factor in determining the interface shear strength between granular soil and geosynthetics. However, these tests can be time-consuming and expensive and may only sometimes be feasible in project planning. Therefore, finding alternative methods to obtain the necessary information is essential. One possible solution is to use reference results from other research. In this way, a database of previous results can be compiled, and a predictive model can be created to estimate the required interface strength values. This study aims to assess the effectiveness of using an Artificial Neural Network (ANN) methodology to predict the shear strength at the interface of sand and geomembrane. A Multi-Layer Perceptron (MLP) architecture was chosen to configure the ANN models, and the training process is a supervised one that involves a Back-Propagation (BP) training algorithm coupled with the Differential Evolution (DE) optimization algorithm. The input data for the models were defined from 428 laboratory tests reported in previous investigations, including 14 input parameters and sand-geomembrane interface strength results. Four ANN models were analysed and compared, differentiated in terms of their number of inputs (9 or 14) and the number of hidden layers (1 or 2). The ANN model with the architecture 14-71-342-1 displayed the most satisfactory results for the training and testing phase in terms of the predicted values' distribution compared to the trend line (R²: 0.919 training, R²: 0.852 testing), a lower number of residual values outside the acceptable range (4% training, 11.6% testing), and excellent prediction performance according to statistical metrics for both phases (RMSE: 1.92, MAE: 1.32, MAPE: 5.03% training, RMSE: 0.852, MAE: 3.36, MAPE: 7.13% training). Based on the results, the ANN technique can be defined as an effective approach for predicting sand-geomembrane interface strength values (friction angle) for the collected data.

This dissertation addressed the study of the filtering behavior of non-woven geotextiles in internally unstable soil under confinement. Permeability tests were carried out under constant head conditions using a permeameter equipped with a device for the application of vertical stresses. The non-woven geotextiles tested had masses per unit area of 200, 400 and 600 g/m², the soil was composed of sand plus different percentages (0, 10, 20 and 40%) of fine material and the applied vertical stresses were 0, 25 and 100 kPa. The drainage materials under the geosynthetic filter were steel spheres of 15 mm in diameter, with ratios between spacing (s) and sphere diameter (d) equal to 1 or 2, or just a perforated plate. A series of filtration tests was conducted under hydraulic gradients of 1 and 10. The results were discussed in terms of the permeability coefficient of the soil-geotextile system (ksg). At the end of the tests, the grain size distribution curves of the material at the base of the soil sample and of the material that impregnated or passed through the geotextile filter were obtained using a laser grain size analyser. The degrees of impregnation of the geotextiles were obtained after the permeability tests. In addition, permeability tests were also performed on the 200 g/m² geotextile after filtration tests. Tests with samples of sand plus 40% of fines with 5, 10 and 15 cm height were performed in order to evaluate whether the height of the specimen influenced the filter behavior. Sagging of the geotextiles in the void spaces between the spheres were also assessed. It was observed that the vertical stress at the top of the samples,the hydraulic gradient, the increase in the mass per unit area of the geotextile and the tests with only the perforated plate under the geosynthetic tended to reduce the permeability of the soil-geotextile system. Increasing the spacing between the spheres decreased ksg and favoured the formation of a blinding layer on the filter surface.

In the present work, grain-scale flow analyses in porous media are presented using hybrid numerical methods. Porous media made from virtual particles with realistic morphology are developed based on the characterization of real granular materials by digital image processing. After morphologically describing a large number of particles, a virtual repository is created with three-dimensional solids that statistically represent the shape of the real particles. The first of the numerical methods used is the algorithm for packing random particles from the repository that generates the virtual porous media. The algorithm is based on the discrete element method (DEM). The modeling of flow within the void matrix obtained from the virtual particle packings is then performed by a computational fluid dynamics (CFD) software. This eliminates the need to use a continuum approach for the porous medium. The methodology presented makes it possible to obtain the distributions of the characteristic flow variables throughout the porous matrix, rather than averaged values obtained for a representative elementary volume (REV). Six packings of virtual particles with different grain sizes were used, with dimensions varying between 4.8 and 32 mm. The results obtained here were then compared with the calculations from flow resistance models widespread in the literature, that were developed with the assumption of the porous medium as a continuous material. Furthermore, the numerical results were validated by experimental data obtained from laboratory tests using a hydraulic channel and real granular material used as a basis for assembling the virtual packings. The merge of numerical methods presented in this work proved to be quite effective, with great analytical value, with qualitative and quantitative advantages. The graphical features available in CFD modeling can be used as powerful tools for visualizing a complex phenomenon that occurs inside the pore network of geotechnical granular materials, represented virtually in an innovative realistic way.

Slopes are soil structures present in the landscape, whether naturally or in man-made geotechnical constructions. Slope failure is usually related to the variation in the water content within the soil during hydrological events such as intense precipitations. Thus, the global stability analysis in unsaturated soil slopes is needed. The present study models the transient factor of safety for global stability in homogeneous and unsaturated soil slopes using the Bishop and GLE (General Limit Equilibrium) limit equilibrium methods. Also, a semi-analytical solution for the transient two-dimensional water flow applied to an unsaturated shear strength model was coupled to the limit equilibrium modeling. This model was implemented in the Wolfram programming language in the Mathematica software. It allowed parametric analyses for the factor of safety varying with time during water infiltration or drying events, applying different boundary conditions for the water content and the slope geometry. This model also allowed the analyses of other state variables such as suction and shear strength. The results of the present model were compared to the results obtained by the same inputs in the commercial software package GeoStudio (Seep/W e Slope/W), presenting consistency amongst themselves for the adopted hypotheses. Some differences between the software for the formulation and the results were explained. The computational tool developed in the Mathematica can be applied in the geotechnical engineering practice for case studies in unsaturated soil slopes, and it may be improved for more complex modeling hypotheses, for both water flow and stability.

The demand for geospatial data today has grown exponentially and given the multiplicity of existing geotechnologies in the market, the production and distribution of data become more agile every day. The interest in georeferenced and structured geotechnical data tends to grow in Brazil due to the mandatory use of the Building Information Modelling (BIM) in the direct or indirect execution of engineering projects and services performed by the entities of the federal public administration or for the manufacture of cartographic products established in the National Policy of Protection and Civil Defense (PNPDEC). Regardless of the purpose of the geotechnical data produced, there is an absence of a storage pattern, a format of integration of the data by various sources and a quality control. Data produced is restricted to information producers causing waste of resources by thoroughly reinvestigating areas that could only be subjected to complementary investigation campaigns. If the management of geospatial data were done adequately, the time used in the compilation and compatibilization of pre-existing data could be invested in the planning of data acquisition, in the production and analysis of derived information. However, for pre-existing data to be used properly, together with the data produced, both need to follow standards and technical specifications that ensure interoperability, sharing, and dissemination. In this context, this research analyzed the relationship between different types of investigations and geotechnical tests and proposed a geospatial data model that defines relationship rules and three-dimensional geotechnical data storage patterns in the database. The rules and standards were presented in the form of conceptual schemes using the Object Modeling Technique for Geographic Applications (OMT-G), because it is a model compatible with the National Spatial Data Infrastructure (SDI). The OMT-G model proved to be adequate to express the relationships and three-dimensional representations of the geotechnical investigations and laboratory tests, although the model is commonly applied to the modeling of two-dimensional geographic data. For the validation of the proposal, a geotechnical database of the Federal District was implemented, using the proposed geotechnical data model, in the PostgreSQL Relational Object Database Manager System (RDBMS) spatially extended with PostGIS, which allowed the evaluation of the use of the data model in practical cases. Based on the geotechnical database, three case studies were conducted addressing data analysis and spatialization in the Olhos D'Água Stream sub-basin, construction of three-dimensional geometry and data selection criteria in the Taquari Housing Sector – Step 2 and analysis of geotechnical characteristics in terrain units defined based on geology, geomorphology and pedology. The entire geotechnical database structure is compiled into an extension to PosgreSQL titled "pggeotec" and available in the Github repository, aiming to stimulate discussion about storage standard and geotechnical data distribution architecture.

In recent years there has been a marked increase in the use of tunnels built on soft soils, especially in urban regions, to serve a variety of purposes, such as transportation (road, rail and subway), and as part of sewage collection systems (Kochen, 2005).
Many cities are developing on soft soils: Shanghai, Bangkok, Kuala Lumpur, Jakarta, Singapore, Bogota, Mexico City, among others. These cities have been presenting related phenomena such as the lowering of water levels in the soil that change the working conditions of underground structures (Rodríguez, 2016).
Measurement of pressure in tunnel lining built in saturated clays in London, Chicago, and Detroit (Peck 1969; Tchebotarioff 1979) shows that vertical pressure evolves over time, sometimes exceeding the initial value of total ground pressure at tunnel axis level, γ*H0. An extreme case of this evolution corresponds to the soft clays subjected to a densification process induced by the abatement of the original hydrostatic pressure in the aquifers that underlie the lacustrine clays, as in the valley of Mexico; in this case, as the piezometric abatement increases, the top of the tunnel liner is subjected to load increases, and at the same time, the sides of the tunnel liner experience a loss of confinement due to reduced water pressure. Pressure differentiation induces important stresses on flexion and compression in the lining (Tamez et al., 1997; Rodríguez et al., 2013).
In this work we study the soil-lining interaction of tunnels built in soft soils subjected to a consolidation process through physical modeling in geotechnical centrifuge, which allows to reproduce the field stresses in a model in a reduced version of the prototype and to accelerate the effects of consolidation phenomena, allowing a considerable reduction in the time scale in relation to the prototype. To obtain long-term stress distribution acting on the support when subjected to piezometric abatement, eight tension cells and five piezometers were installed in the tunnel lining, as well as 16 strain gauges to obtain bending moments and axial forces. In addition, five piezometers were placed in the soil mass at different depths to obtain the magnitude of the piezometric abatement and three surface extensometers to obtain the surface settlements.

The construction of structures to protect against floods, water diversions, breaking waves through conventional construction methods has led to an increase in the use of alternative methods, including the use of geotextile tubes. This method has quick implementation as well as reduced costs. Other uses with of geotextile tubes have been directed to the dewatering of fine-grained materials resulting from industrial processes or wastewater treatment with different concentrations. However, few studies have been carried out to understand the behavior of these structures in terms of the type of geotextile and filling material and technique. In this sense, the present work aimed to study the behavior of geotextile tubes used in environmental protection works. For this purpose, tests were defined using sandy and clayey soils as filling materials and non-woven geotextile in the manufacture of tubes. Thus, filling of single and stacked tubes conditions were carried out. The tests were instrumented in order to obtain variations in the height of the tubes, pore pressures in the filling material as well as the vertical stresses at the base of the tubes. Variations in the deformation of the geotextile were obtained during the test and samples of the soil particles passing through the geotextile were collected for the study of its retention capacity. In addition, vane tests were performed on the filling material to obtain its undrained strength. Results for tubes filled with sand allowed analyzing the rapid stabilization process of the dewatering process, as well as the stabilization of stresses and pore pressures at the base of the tube and served as a reference to calibrate the system. After this process, the clayey soils were used to fill single and stacked tubes. The results obtained showed that the initial dewatering process was characterized by pore pressure peaks and total stresses at the base of the tubes that stabilized over time. The greatest deformations in the geotextile were obtained at the end of the filling process with small variations during the test. The variation of undrained strength with depth in the stacked tubes and in two of the single tubes showed a decrease in strength with depth due to the impermeable condition of the foundation. All these results showed the importance of verifying the behaviour of the nonwoven geotextile both for isolated and stacked tubes, which can modify the geometry and influence the dewatering process, the strength of the filling material and the strains in the geotextile.

Brazil is predominantly covered by soils with tropical characteristics. The state of Mato Grosso currently stands out as the country's main grain producer. The production, which requires storage, and the timely marketing of the crop have encouraged investments in grain storage facilities on farms. In addition, the state is expanding its road and rail network, consequently requiring foundations that meet the technical, environmental, and economic needs of these significant projects. Soil improvement techniques are seldom used in Brazil; however, outside the country, they have been widely employed in various forms and for different soil types, representing a technically and economically effective alternative. This research evaluates the technical viability of soil improvement for use in foundations set on unsaturated lateritic soil, with the addition of Portland cement using the Deep Mixing Method (DMM). The study area is located in the municipality of Primavera do Leste, MT. The soil in the study area was characterized with laboratory and field tests. The soil-cement mixture was initially analyzed as per the recommendation of the Swedish Geotechnical Society. The parameters consolidated in the literature are mostly applied to soft soils and organic soils. Therefore, to understand the behavior of the DMM in lateritic soils, 14 DMM piles were executed in the study area, with a diameter of 0.6 m and a depth of 3.5 m. It was concluded that if the geotechnical characteristics of the soil are evaluated before applying the DMM to lateritic soils and a careful study of the mixture with the soil is carried out, a good material for the piles is guaranteed. With the addition of 120 kg/m³ of Portland cement, an average reduced rupture stress through the global safety factor of 2 (σ50) of 500 kPa and a deformation modulus (E) of 25 MPa were obtained. These values indicate the possibility of applying the technique. In addition, environmental benefits are obtained such as reducing the exploitation of aggregates and reducing the emission of CO2 in the manufacture of cement, due to the reduced amount of binder per m³ of material.

Civil construction serves as a cornerstone for economic and societal progress within nations. However, its undeniable significance is accompanied by a concerning ecological footprint. Recent studies have indicated that the construction sector accounts for nearly half of global natural resource consumption and is responsible for a quarter of waste generation and CO2 emissions. This predicament underscores a pressing environmental challenge. In response, the recycling and repurposing of Construction and Demolition Waste (CDW) within the construction industry presents a potent avenue to safeguard natural resources, mitigate environmental impacts, and foster sustainable development. In this context, this study aims to assess the viability of employing CDW as structural fill material in mechanically stabilized earth walls through full-scale pull-out tests. To gauge its efficacy, the findings are juxtaposed against the geomechanical behavior of conventional natural sand, widely utilized in reinforced soil solutions. Furthermore, design parameters derived from laboratory tests, accounting for diverse design conditions, are compared against values established by standard technical protocols and design guidelines. Furthermore, a meticulous analysis of chemical and electrochemical attributes is conducted to evaluate durability and resistance to corrosion and degradation of reinforcement elements. Despite not fully conforming to the grain size distribution criteria stipulated by several international regulations – which advocate for the use of well-graded soils – the recycled CDW demonstrates commendable performance and remarkable mechanical characteristics, particularly regarding the resistance to pullout of metallic reinforcements with bumps and conventional synthetics. This performance level is on par with or surpasses that of natural sand, exhibiting comparable or superior pull-out resistance parameters. Consequently, the application of recycled CDW as an environmentally friendly alternative for total or partial replacement of conventional natural aggregates emerges as a promising solution. The viability of this approach is contingent upon satisfying additional design prerequisites. It is noteworthy that the recycled CDW gravel exhibits elevated concentrations of chloride and sulfate ions, alongside diminished electrical resistivity values. Complementing these findings, an analytical model is formulated based on pull-out tests, capable of predicting the response of metallic ribbed strips in alternative geotechnical materials, predicated upon their grain size distribution. This model adds a valuable predictive dimension to the utilization of sustainable materials in geotechnical engineering applications.

With the increasing demand for underground construction projects, there has been a growing need to understand and predict the behavior of rocks when subjected to different types of external loads. Multiple experimental and numerical techniques have been developed to represent the various phenomena involving rocks and rock masses. One of these techniques is the Continuum Voronoi Block Model (CVBM), which, with its pseudo-discontinuous approach, has proven to be a numerical tool capable of simulating rock behavior on both field and laboratory scales, assuming a single material for the entire set of blocks and joints used in its representation. However, in recent years, the contribution of other sources of heterogeneity for a more accurate modeling of rocks has been recognized, allowing for more realistic failure modes and a development of macroscopic stresses closer to laboratory results, especially in cases of confined compression. For these reasons, in this work, material and contact heterogeneity were integrated into the CVBM using a script developed in the Python language. As a case study, the behavior of Äspö diorite was represented when subjected to different types of laboratory tests (simple and confined compression, indirect tension, cyclic loading, and fracture toughness), analyzing the impact of different parameters, the size of constituent elements, and spatial and structural variability on elastic response, characteristic stresses, and model strengths. The heterogeneous CVBM adequately simulated the rock's macroscopic response, capturing the natural variability of the results and the development of different types of fractures throughout the loading process. The new features added to the model resulted in early development of tensile stresses within the material, causing the gradual growth of fractures, impacting characteristic stresses and the extent of the unstable fracture growth zone. Additionally, other types of rocks with specific internal structural characteristics (foliation planes or inclusions) were represented, culminating in the extension of the procedure to the application of the probabilistic approach for defining properties within the test body.

The technique of rigid inclusions has been commonly used to reinforce soft soils of road and railway embankments. However, nowadays it has been used to reduce costs of building foundations. A numerical study developed by Chevalier et. al. (2011) shows that the load transfer mechanism is different when a rigid slab is used instead of an embankment and its efficiency is considerably higher. This paper presents the results of a numerical-experimental study of the load transfer mechanism between the head of an inclusion and the distribution layer under a rigid slab, considering a cohesive-granular soil formed by a tropical soil characteristic of the Federal District. The numerical models were calibrated and validated using data from simplified, full-scale physical models. The development of the numerical model with discrete elements method (DEM) aimed to simulate and validate the development of the load transfer mechanism observed in the physical model. From the definition of this mechanism, a finite element method (FEM) model was developed, and it allowed simulating the load-transfer platform´s (LTP) behavior observed in the physical and DEM models. After calibration and adjustment of the strength and compressibility parameters of the materials used in the LTP from laboratory tests and the physical models to the Hardening Soil (HSM) constitutive model, probable rupture mechanisms in the LTP were studied using the models based on the FEM and the calibrated and adjusted parameters for the HSM, observing the influence of the shear strength parameters of the LTP and the geometry of the load transfer cone on the rupture mechanism. With the results observed, a methodology was developed for obtaining the load capacity of the inclusion head that allows the definition of the maximum load that can be transferred by the load-transfer cone, which allows obtaining the maximum CDC thickness and the minimum spacing between inclusions.

Horizontal drains are frequently used as a mean to reduce porepressure and improve block stability in fractured rock masses. Given the difficulties in data collection, discrete fracture networks (DFNs) are a useful tool for water flow studies, as they enable the stochastic representation of discontinuities and the generation of many models. The objective of this work was to calculate water flow in DFNs and to provide guidance for the design of more efficient drainage systems. A computational code was developed for this purpose, based on equilibrium at the nodes, through saturated meshes created with the UnBlocksGen program. Different values for mean and standard deviations of hydraulic aperture were tested for the Monte Seco rock mass, and water flow was simulated with three different hydraulic aperture hypotheses: constant, variable and correlated with the discontinuities’ length. Water flow was then calculated for the Yangfanggou, El Teniente and Monte Seco rock masses using the method of Reeves et al. (2013) with constant and variable hydraulic apertures, to determine drain length and orientation. The efficiency of drainage was tested for the Monte Seco rock mass on the basis of drain length, orientation, and quantity. The results showed that water flow increased significantly with mean hydraulic aperture, but decreased for higher values of the standard deviation of hydraulic aperture. Water flow through the meshes was greater for hydraulic aperture correlated with discontinuity length, followed by constant hydraulic aperture and, lastly, variable hydraulic aperture. When hydraulic aperture was used instead of transmissivity with the method of Reeves et al. (2013) results were strongly influenced by boundary conditions, making the determination of drain length more difficult. Drainage efficiency was greater for drains orthogonal to the rock face and for longer drains. Furthermore, fewer but longer drains were more efficient than a larger number of shorter drains. The results obtained may help to better design water drainage systems. 

 Evaluation of soil-geosynthetic interaction is important for analyzing the stability of the overall structure. This is because the interaction between the reinforcing geosynthetics and the reinforced soil can be a start for breakage which may cause a structural failure. Several researchers have been studying factors determining the interface shear strength in the laboratory and identified various components which affect the overall strength outcome. Machine Learning has thatgreat potential for the analysis of parameters which are influenced by many variables. This dissertation brings a discussion about the use of Random Forest regression, which is a Machine Learning algorithm, for predicting geomembrane-sand interface friction angle.  

 The interfaces subjected for strength parameter investigation include geomembrane and cohesionless soil, and 495 interfaces from various literature were collected. The acquired interface data is utilized for the overall statistical and Machine Learning analysis. Fourteen parameters were recorded from the referred literatures as the factors determining the interface shear strength. The fourteen parameters are from three main interface components which are; laboratory test type, geomembrane properties and soil properties.  

 The presented data has been studied by using simple linear regression before initializing the Random Forest, to evaluate the interdependence between pairs of influencing parameters and their correlation with interface friction angle. The Pearson's correlation coefficient results are indicating, the influence level between the interface components is mostly not strong. These correlation values imply the nonlinear distribution of the database and the importance of a multivariate and nonlinear algorithm for studying the referred types of interfaces.  

 After the data analysis, an optimized Random Forest has been initialized to predict interface friction angle. It is observed only for 3% of the training set and 6% of the testing set that the friction angle estimation has exceeded ±5° from the laboratory records. The coefficient of determination measures shows strong coherence between friction angles from laboratory studies and Random Forest estimations by resulting R² = 0.93 and R² = 0.92; for the training and testing sets respectively.  Thus, the Random Forest has forecasted interface friction angle adequately

Geosynthetics have proven to be beneficial in reinforcing soils, especially in problems with large deformations, such as unpaved roads built on soft soils and subjected to high loads. Due to heavy machinery traffic, this situation occurs even during the construction period of the road, in which is necessary to perform surface maintenance for the execution of the fill layers. Several experimental studies have indicated an improvement in the mechanical performance of unpaved roads when subjected to surface maintenance. In this context, the present research aimed at studying, by means of the Finite Element Method, the behavior of unreinforced and reinforced unpaved roads subjected to surface maintenance. The analyses consisted of the following steps: determination of geostatic stresses in the subgrade; inclusion of the fill layer, as well as the geosynthetic reinforcement at the interface between the materials; application of a distributed load on the fill; unloading; execution of the surface maintenance from the deformed configuration of the fill; and reapplication of the load. The developed methodology was validated by comparisons with experimental results presented in the literature, indicating that these constructive steps allow a more realistic simulation of the mechanical behavior. The comparisons were made in terms of stress and displacement of the system at the base of the loading plate and the general behavior of total displacements, which reflect the influence of surface maintenance and geosynthetic reinforcement on the bearing capacity of the system and the observed failure mechanisms. Furthermore, relevant information was obtained as shear stresses mobilized in the system and deformation in the reinforcement. This was also studied by comparisons with data predicted empirically and by parametric analyses whose variables were height and stiffness of the embankment layer.

• The studies of the mechanical parameters of granular materials have been, simply, based on the continuum. The simplicity of studying the environment through phenomenological constitutive models and representative volume tests, made it impossible to analyze the characteristics of the samples at the particulate level. With the development of new tools such as supercomputers and image analyzers, the discrete element method, which is based on the interaction of grains, has been gaining strength. In this approach it is possible to consider the microscopic characteristics of the particles in the evaluation of the mechanical behavior of granular materials. In this work, the influence of size, shape and surface texture, at the particle level, on the angles of repose and friction of granules is studied. For this, granulometric and morphological characterization tests were carried out, evidencing parameters such as D50, sphericity, circularity, roundness. Through the tests of angle of repose, direct shear and saturated CD triaxial, the mentioned angles were determined. The experimental results obtained showed the same trend of behavior of the materials in relation to the resistant angles. The characterization of the particles showed that manufactured materials, spheres and polyhedrons, are similar in relation to surface texture while the sands are distinguished in the sample group itself. The sizes were different as well as the shape indices, although the spheres had similar indices. In this research, the size parameters did not show a good trend in relation to the angles. Among the shape indices, the roundness had the best coefficient of determination. Using the roundness index, equations from the literature were selected that had it as a variable. In addition to tests with existing equations, equations with one and two variables were also proposed for each type of angle studied. The chosen parameters were the roundness, since it was better related to the angles, and the D50 (for friction in direct shear), eMÁX (angle of repose) and D10 (friction in the saturated CD triaxial), in the functions of two variables. Finally, statistical analyzes were performed with the equations from the literature and the proposals verifying that the equations of this study had a better index of precision and accuracy.

The consideration of soil in the unsaturated condition has become increasingly common and necessary in geotechnical problems. As for soil modeling in this condition, it is known that the progress in its use is linked to the in-depth studies and also to computational advances. When considering the volumetric variation of a soil, the soil water retention curve and the k-function become changeable hydraulic properties and, with this, the concepts of Soil–Water Retention Surface (SWRS) and unsaturated hydraulic conductivity surface are established. This thesis aims to propose new constitutive models of the mentioned surfaces for unimodal and bimodal behavior soils. The development starts from an existing unimodal model of soil water retention curve and the bimodal model is elaborated from the linear superposition principle. Its experimental validation was performed and, with this, a simulation of the water flow in this soil was made by finding and solving the partial differential equation governing the phenomenon. The relationship between the uni and bimodal models and the porosimetry was also made, where the curve models of the cumulative pore–size distribution and the direct pore–size distribution were created and validated, which, as a consequence, consolidate the physical meaning of some of the parameters used. Finally, the SWRS models for uni and bimodal soils are created and also validated with experimental data, and during this development, the physical meaning of the hydraulic parameter d is consolidated and an analytical expression for finding the air–entry value of a soil is also developed. An analytical formulation for the developed concept of hydraulic conductivity surface for uni and bimodal soils was proposed and its three-dimensional graphical representation was made. During the development of the thesis, an attempt was made to present a coherent mathematical rigor and to represent, with accuracy, the behavior of the soils under the established conditions.

The advances in equipment and tools implemented for geotechnical engineering are allowing
projects to be supported by numerical simulations. However, even with sophisticated resources,
the industry still lacks little information on laboratory parameters as in the case of Concrete
Face Rockfill Dams (CFRD). In this sense, the present thesis becomes another academic
provocation to stimulate the use of backanalysis of real BEFC works to search for simplified
models that can be applied in the industry. The objectives include the simplification of the
behaviour for linear modules associated with a coefficient representing the physical
characteristics of settlements in time and due to reservoir flow. The calibrated model allows
three-dimensional extrapolation for the evaluation of stresses and displacements in the concrete
slab at different construction delays. Among several topics, the text goes through examples of
real BEFC simulations and cites some constitutive models of collapse and flow in geotechnics.
The model for calibration of the BEFC includes elegant adaptation of the heat flux as a
representation artifice of the vertical orthotropic collapse of the rockfill. For calibration, 16
constitutive models were tested, nine of them got well-fit results both in the constructive phase
and during backfilling, compared to read displacement of BEFC. After collecting the results,
the parameter pairs were plotted in point cloud in search of patterns for the calibrations. Further,
six different slab constructive delay sequencing arrangements were simulated. The study uses
a numerical artifice that enables the activation of the concrete slab in its correct design
thickness. Among the results of stresses, the step-by-step sequencing of the slab following the
rockfill reached the worst horizontal stresses, which reach up to 48 MPa, which exceeds the
typical compressive strengths of conventional concrete and becomes a point of attention
regarding the structural integrity of the slab. The results suggest that step-by-step sequencing
should be avoided and a balance between behavior could be achieved in a scenario of delay
sequencing of the slab concreting of at least half the height of the dam with initial backfill. It is
important to portray that all the considerations in the thesis are for a specific example. Any
generalization should be avoided or be associated with larger complementary studies, because
each dam is a unique project and should be studied individually, on a case by case basis.

In this work, a 3D numerical model was developed in a software based on the Finite
Element Method, using the results obtained in a geotechnical centrifuge model of a
piled raft system founded on soil subjected to regional subsidence. The numerical model

was calibrated for an initial configuration of 9 piles distributed in the center of the raft.
Soil parameters were obtained, calibrated and validated for the Hardening Soil Model
based on laboratory results of triaxial and consolidation tests. In addition, other
laboratory tests were carried out in a centrifuge that allowed obtaining the resistance
parameters of the foundation studied.
The developed numerical model reproduced satisfactorily soil and foundation
consolidation displacements due, not only by the structural service load but also by the
pore pressure drawdown. For load distribution on piles and raft, the model reproduces
with good agreement the foundation behavior only for the structural service load, for
pore pressure drawdown some adjustments on the embedded piles elements shaft and
base resistance had to be done. The developed model allowed to identify the most
sensitive parameters for this type of simulation, to define the types and stages of
analysis that had the best fit for the physical model, and to obtain additional results to
those measured in the physical model, e.g., the axial load distribution developed along
the piles and therefore the magnitude of the negative skin friction, that is an important
load that should be considered for the structural safety review of piled foundations
subjected to this complex conditions.
This model was used to carry out a parametric analysis that allowed evaluating the
influence of geometric characteristics such as the spacing, length and slenderness of the
piles in the piled raft systems. With the assessment of these results, some correlations
are presented to estimate the negative friction from these characteristics.

Uma parte significativa dos reservatórios geotérmicos e de petróleo e gás possuem
fraturas naturais que impactam sua performance. Quando essas discontituidades se
encontram na escala sub-sísmica, sua incorporação aos modelos numéricos é desafiadora,
pois os custos computacionais de sua representação explícita são geralmente proibitivos.
As soluções mais populares que consideram o efeito dessas fraturas de pequena escala
são os modelos de dupla porosidade e o cálculo de propriedades equivalentes (upscaling).
No entanto, enquanto os modelos de dupla porosidade consideram geometrias muito
idealizadas e pouco representativas de redes de fraturas reais, as técnicas tradicionais de
upscaling não são capazes de capturar a influência dinâmica das fraturas, cujas
permeabilidades mudam continuamente durante a vida produtiva do reservatório. Esta
tese desenvolve métodos e ferramentas computacionais para a modelização multiescala
do comportamento hidro-mecânico de reservatórios contendo redes complexas de
fraturas. O método multiescala adotado é uma adaptação do Médoto dos Elementos
Finitos (MEF) multi-nível, em que a microescala e a macroescala são resolvidas
simultaneamente com o FEM e acopladas de acordo com os princípios da homogenização.
A modificação aqui proposta é denominada método Box multi-nível, pois o MEF foi
substituído pelo método Box. Ao contrário das técnicas convencionais de upscaling, este
método captura os efeitos dinâmicos das heterogeneidades sem a necessidade de definir

modelos constitutivos para a macroescala. No nível do Volume Elementar Representativo
(VER), as fraturas são geradas de maneira estocástica e representadas por elementos de
interface. Um programa de código aberto foi estendido para comportar simulações
hidromecânicas em meios fraturados elastoplásticos. Uma nova metodologia estatística
baseada no Teorema do Limite Central para definir o tamanho do VER de meios
fraturados estocásticos foi proposta. Além disso, dois métodos para a imposição de
condições de contorno periódicas foram adaptados para meios contendo elementos de
interface. Os métodos e ferramentas desenvolvidos foram aplicados em um caso sintético
de depleção de um reservatório inspirado em um carbonato fraturado real. O método
multiescala foi capaz de representar a perda de produtividade causada pelo fechamento
das fraturas e a evolução anisotrópica dos campos de poropressão.