Solos que contem sulfatos apresentam complicações no seu comportamento quando são tratados com estabilizadores à base de cálcio como a cal. Quando um solo que contem sulfatos reage com cal, formam-se minerais expansivos como a etringita e a taumasita que são responsáveis pela deterioração e falha de vários projetos de solos estabilizados. Tem-se bem demostrado sob ensaios de laboratório que a relação vazios/(agente cimentante) é um parâmetro adequado para a avaliação e previsão de comportamentos mecânicos, como a resistência à compressão simples de vários tipos de solos cimentados artificialmente. Este trabalho pretende encontrar se as previsões da relação vazios/cal na resistência à compressão simples são efetivas para dimensionar misturas de solos sulfatados estabilizados com cal construídas em campo. Para lograr esse objetivo executou-se um programa experimental de ensaios de laboratório em conjunto com a construção de trechos experimentais de solo-cal. Os ensaios de laboratório ajudaram comprender o comportamento de solos sulfatados estabilizados com cal, e demostraram que a relação vazios/cal controla a resistência à compressão simples desses solos estabilizados evidenciando uma relação coerente com os resultados das resistências de campo, encontrando-se que as resistências de campo e laboratório são controladas por essa relação e que podem ser previstas pela mesma Ensaios de Difração de Raios X (DRX) e Microscopia Eletrônica de Varredura (MEV) identificaram minerais expansivos (etringita) nas amostras estabilizadas com cal e cinza-cal. Finalmente em laboratório foram testadas algumas soluções recomendadas na literatura para melhorar a estabilização de solos sulfatados mediante um projeto experimental demostrando que a estabilização com cinza volante e cal melhoram consideravelmente a resistência, a durabilidade e a estabilidade volumétrica de solos sulfatados enquanto que a técnica do mellowing mostrou melhoras apenas na estabilidade volumétrica do material. / Sulfate rich soils present complications in their behavior when are treated with calcium-based stabilizers such as lime. When a soil containing sulfates reacts with lime, expansive minerals such as ettringite and thaumasite are formed which are responsible for the deterioration and failure of various stabilized soil projects. It has been well demonstrated under laboratory tests that the relationship porosity/cementing agent ratio is an appropriate parameter for the evaluation and prediction of mechanical behavior, such as the unconfined compression strength of various types of artificially cemented soils. This work intends to find out if the predictions of the relation porosity/lime in the unconfined compression strength are effective to design lime stabilized soils mixtures built in field. In order to achieve this objective an experimental program of laboratory tests was carried out together with experimental road sections of lime stabilized soils were built. The laboratory tests helped to understand the behavior of lime stabilized sulfate soils, and demonstrated that the void/lime ratio controls the unconfined compression strength of these stabilized soils and that exist a consistent relationship with the field unconfined compression strength results, finding that both, field and laboratory strengths are controlled and can be predicted by this ratio X-ray diffraction (XRD) and Scanning Electron Microscopy (SEM) tests were performed to identify expansive minerals (ettringite) on soils samples stabilized whit lime and fly ash-lime. Finally, some solutions recommended in the literature have been tested in laboratory to improve the stabilization of sulfate rich soils through an experimental design, showing that the fly ash-lime stabilization improve considerably the strength, durability and volumetric stability of the sulfate rich soils whereas that for the mellowing the improve was observed only on the volumetric stability of the material.

In developing a constitutive model that could predict the settlement due to the collapse, several goals needed to be met. These were to gain an understanding of the collapse phenomenon, knowing the soil properties at the natural water content and how they change after collapse, and develop and test the new model. It was felt that laboratory testing could be of use. The types of test conducted included use of the Oedometer, Pressuremeter, and Triaxial tests. The material that was used for the testing was a "generic" soil manufactured out of diatomite. In all of the tests the soil was tested dry and saturated in order to establish state limits of the soil. Next, the soil was loaded dry then inundated which initiated the collapse of the soil. The stress and strains were continually recorded. From the testing it was concluded that there is a stress-strain region where after collapse the soil looses considerable strength. With increasing stress and strain the soil eventually becomes stronger. From the triaxial tests, the stress-strain data from this "region of collapse" was used in a constitutive model. Stress paths from the Oedometer and Pressuremeter tests were then successfully applied to the model. The constitutive model used was an elasto plastic model. The elastic and plastic strain components were provided using functions for yielding, hardening, plastic potential, and failure as proposed by Paul Lade in his work on cohesionless, frictional materials. Results from the conventional triaxial shear tests and isotropic compression tests were used to derive the values of the functions for the model. The end result was three dimensional surfaces for failure, yielding, plastic work and plastic potential for the dry and saturated soil in the zone of collapse.

Forest soils with low bulk densities are often considered less susceptible to compaction than soils with higher bulk densities. The objective of this study was to determine if soil strength controlled the compression of soils with low bulk density. Four soils were selected for this evaluation. Three of these were andic soils with low bulk density and the fourth soil was a more dense, cohesive soil. Undisturbed samples of saturated and partly saturated soil were compressed in a one-dimensional consolidation test apparatus. Measurements with separate samples were at one of 7 normal stresses between 0.033 and 1.96 MPa. Shear strength of saturated soil was measured in direct shear tests. Primary consolidation of saturated soil was completed in less than one minute at all normal stresses. Shear stress and bulk density increased continuously during shear strain. The compression index of the cohesive soil was significantly larger (p<0.05) than that of the andic soils. The shear strength of andic soils (average cohesion intercept of 0.016 MPa and friction angle of 33.3°) was significantly higher (p<0.05) than the cohesive soil (cohesion intercept of 0.028 MPa and friction angle of 28.9°). When saturated, the cohesive soil was more compressible than the andic soils because of lower soil strength. A nonlinear model of soil compression was developed that accurately predicted the compressed density of saturated and partly saturated soil as a function of normal stress, initial bulk density of undisturbed samples, and degree of saturation. As degree of saturation decreased, the compressibility of the cohesive soil decreased more rapidly than it did for the andic soils. As a result, bulk density of dry cohesive soil increased less than it did for dry andic soils. Differences in the compressibility of soils were attributed to texture and clay mineralogy. The differences in the compressibility of these soils were much smaller than were the differences in bulk density. Decreasing water content affected the compressibility of the cohesive soil more than it affected the andic soils. Because soil strength controls the compressibility of these forest soils regardless of bulk density, it will also determine the susceptibility of soils to compaction by machines. / Graduation date: 1991

Geotechnical engineers often face important discrepancies between the observed and the predicted behavior of geosystems. Two conceptual frameworks are hypothesized as possible causes: the ubiquitous spatial variability in soil properties and process-dependent terminal densities inherent to granular materials. The effects of spatial variability are explored within conduction and diffusion processes. Mixtures, layered systems, inclusions and random fields are considered, using numerical, experimental and analytical methods. Results include effective medium parameters and convenient design and analysis tools for various common engineering cases. In addition, the implications of spatial variability on inverse problems in diffusion are numerically explored for the common case of layered media. The second hypothesis states that there exists a unique terminal density for every granular material and every process. Common geotechnical properties are readily cast in this framework, and new experimental data are presented to further explore its implications. Finally, an unprecedented field study of blast densification is documented. It involves comprehensive laboratory and site characterization programs and an extensive field monitoring component. This full scale test lasts one year and includes four blasting events.

Expansive soils are a worldwide problem that poses several challenges for civil engineers. Such soils swell when given an access to water and shrink when they dry out. The most common and economical method for stabilizing these soils is using admixtures that prevent volume changes. In this study the effect of using Bigadic zeolite (boron by-product) in reducing the swelling potential is examined. The expansive soil is prepared in the laboratory by mixturing kaolinite and bentonite. Bigadic zeolite (boron by-product) is added to the soil at 0 to 25 percent by weight. Grain size distribution, Atterberg limits and swell percent and rate of swell of the mixtures are determined. Specimens are cured for 7 and 28 days. As a result of the experimental study, it was seen that addition of Bigadic zeolite (boronby-product) decreased swelling potential and rate of swell of the artificially prepared expansive soil specimen at laboratory conditions. The swell percentage and rate of swell of the stabilized specimens are affected positively by curing.

Expansive soils are treated with anhydrous or hydrated lime. The use of calcium-based stabilizers such as calcium oxide (lime) in sulfate-bearing clay soils has historically led to distress due to the formation of an expansive mineral called ettringite and possibly another such mineral, thaumasite. Predicting the precipitation of these minerals is a complex problem related not only to soil composition but also construction methods, availability of water, ion migration, and whether the expansive mineral growth can be accommodated by the void structure in the surrounding soil. In trying to control the damage associated with such occurrences, engineers have attempted to determine a threshold value of soluble sulfates, a quantity that is relatively easy and quick to measure, at which significant ettringite growth and, therefore, structural distress occurs. Unfortunately, experience alone and “rules-of-thumb” based on experience are not sufficient to deal with this complex issue. This thesis describes how thermodynamic geochemical models of lime-treated soil can be used as a first step toward establishing problematic threshold levels of soluble sulfates for a specific soil. A foundation for the model development is presented, and two different soils are compared to illustrate their very different sensitivities to ettringite growth upon the addition of hydrated lime. Various soil series along the route of SH 130 between Austin and San Antonio have been identified to contain soluble sulfate that may pose a problem for soil stabilization using lime and cement. Since the model predicts ettringite growth based upon site-specific properties, this thesis also shows how the model can be used to assess the potential amelioration effects of soluble silica. Research was conducted at the Texas Transportation Institute to develop a methodology for identifying areas which are susceptible for ettringite formation. The proposed methodology uses a magnetometer to quickly screen large areas for high sulfate. Application of GIS to identify ettringite formation using soils, topographical, and geological maps is also illustrated in this thesis.

published_or_final_version / abstract / toc / Civil Engineering / Master / Master of Philosophy

published_or_final_version / abstract / toc / Applied Geosciences / Master / Master of Science

The main project was an assessment of the effect that colonization by five different Vesicular-arbuscular mycorrhiza (VAM) cultures have on the growth of the indigenous-grasses Themeda triandra and Trachypogon spicatus, when grown on coalmine topsoil. With unamended topsoil, VAM showed the ability to significantly increase the growth of the grasses compared to non-VAM control plants. The amount of effect varied with the VAM inoculum culture type, with a VAM culture originally from the Cape Flats being the most effective. In a second trial, soil fertilized with nitrogen, potassium and low concentrations of phosphate (P) was used. Again VAM displayed the ability to improve grass plant growth. The increase in P caused the Large spore inoculum to become the most effective. This indicated that different VAM cultures are inhibited to different degrees by an increase in phosphate fertilization. The low level of VAM infection, in both trials, seemed to preclude most of the VAM associated nutrient uptake control. Varying reports have been published on the effect of fertilization on VAM infection and colonization. In an attempt to further elucidate the role of fertilizer in VAM inhibition, rhizosphere soil from a long term fertility trial near Witbank, S.A. was sampled. Amcoal environmental services fertilized forty-two plots with varying concentrations of nitrogen, potassium, phosphate and lime to assess the growth of a variety of grasses. The trial had been maintained for ten years before sampling was completed for this project. Samples from each plot were taken from the rhizosphere soil of the most prominent grass (Digitaria eriantha). VAM spores were extracted from all the samples and five different types of spores were identified and counted for each sample. By comparing spore counts from each plot, the effect that the fertilizer regime had on the VAM on that plot could be assessed. Variation in the concentrations of nitrogen (N) and potassium had no significant effect on VAM colonization. Very low concentrations of N could not be assessed as all plots had been initially top dressed with nitrogen fertilizer. Phosphate (P) fertilizer concentration had a marked effect on spore concentrations. There was a significant increase in spore concentration as P levels were increased from zero P fertilization to 80kgs P/ha. Further increase in P to ≥ 60kgs P/ha resulted in a significant decrease in spore concentrations. From this it would appear that a low level of soil P is needed to give maximum VAM colonization and further increase in soil P causes VAM inhibition. Lime ameliorated the VAM inhibition caused by high concentrations of P. Increase in P caused spore concentrations of low abundance propagules (LAP) too decreased more rapidly than high abundance propagules (HAP). In high P soils VAM with LAP would eventually be eliminated from the system resulting in a decrease in VAM diversity. A project was attempted to use the recently developed Randomly Amplified Polymorphic DNA in conjunction with the Polymerase Chain Reaction (RAPD PCR) techniques to identify different VAM families. The technique causes the amplification of segments of DNA which can be visualized by gel electrophoresis and staining. Band patterns formed can be related to the VAM of origin and hence used in identification of that VAM. An attempt was made to amplify DNA from a single spore in this manner which would, in conjunction with morphological observations, make identification of VAM easier and more accurate. Problems with either releasing the DNA from the spores, or substances in the spore inhibiting the PCR reaction made obtaining band patterns difficult. After many PCR attempts, varying extraction methods and PCR conditions, no repeatable results could be obtained and work on this project was discontinued. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 1997.

River stage or flow rates are required for the design and evaluation of hydraulic structures. Most river reaches are ungauged and a methodology is needed to estimate the stages, or rates of flow, at specific locations in streams where no measurements are available. Flood routing techniques are utilised to estimate the stages, or rates of flow, in order to predict flood wave propagation along river reaches. Models can be developed for gauged catchments and their parameters related to physical characteristics such as slope, reach width, reach length so that the approach can be applied to ungauged catchments in the region. The objective of this study is to assess Muskingum-based methods for flow routing ill ungauged river reaches, both with and without lateral inflows. Using observed data, the model parameters were calibrated to assess performance of the Muskingum flood routing procedures and the Muskingum-Cunge method was then assessed using catchment derived parameters for use in ungauged river reaches. The Muskingum parameters were derived from empirically estimated variables and variables estimated from assumed river cross-sections within the selected river reaches used. Three sub-catchments in the Thukela catchment in KwaZulu-Natal, South Africa were selected for analyses, with river lengths of 4, 21 and 54 km. The slopes of the river reaches and reach lengths were derived from a digital elevation model. Manning roughness coefficients were estimated from field observations. Flow variables such as velocity, hydraulic radius, wetted perimeters, flow depth and top flow width were determined from empirical equations and cross-sections of the selected rivers. Lateral inflows to long river reaches were estimated from the Saint-Venant equation. Observed events were extracted for each sub-catchment to assess the Muskingum-Cunge parameter estimation method and Three-parameter Muskingum method. The extracted events were further analysed using empirically estimated flow variables. The performances of the methods were evaluated by comparing both graphically and statistically the simulated and observed hydrographs. Sensitivity analyses were undertaken using three selected events and a 50% variation in selected input variables was used to identify sensitive variablesThe performance of the calibrated Muskingum-Cunge flood routing method using observed hydrographs displayed acceptable results. Therefore, the Muskingum-Cunge flood routing method was applied in ungauged catchments, with variables estimated empirically. The results obtained shows that the computed outflow hydrographs generated using the Muskingum-Cunge method, with the empirically estimated variables and variables estimated from cross-sections of the selected rivers resulted in reasonably accurate computed outflow hydrographs with respect to peak discharge, timing of peak flow and volume. From this study, it is concluded that the Muskingum-Cunge method can be applied to route floods in ungauged catchments in the Thukela catchment and it is postulated that the method can be used to route floods in other ungauged rivers in South Africa. / Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2005.