abstract: The dissimilatory reduction of nitrate, or denitrification, offers the potential of a sustainable, cost effective method for the non-disruptive mitigation of earthquake-induced soil liquefaction. Worldwide, trillions of dollars of infrastructure are at risk for liquefaction damage in earthquake prone regions. However, most techniques for remediating liquefiable soils are either not applicable to sites near existing infrastructure, or are prohibitively expensive. Recently, laboratory studies have shown the potential for biogeotechnical soil improvement techniques such as microbially induced carbonate precipitation (MICP) to mitigate liquefaction potential in a non-disruptive manner. Multiple microbial processes have been identified for MICP, but only two have been extensively studied. Ureolysis, the most commonly studied process for MICP, has been shown to quickly and efficiently induce carbonate precipitation on particle surfaces and at particle contacts to improve the stiffness, strength, and dilatant behavior of liquefiable soils. However, ureolysis also produces copious amounts of ammonium, a potentially toxic byproduct. The second process studied for MICP, denitrification, has been shown to precipitate carbonate, and hence improve soil properties, much more slowly than ureolysis. However, the byproducts of denitrification, nitrogen and carbon dioxide gas, are non-toxic, and present the added benefit of rapidly desaturating the treated soil. Small amounts of desaturation have been shown to increase the cyclic resistance, and hence the liquefaction resistance, of liquefiable soils. So, denitrification offers the potential to mitigate liquefaction as a two-stage process, with desaturation providing short term mitigation, and MICP providing long term liquefaction resistance. This study presents the results of soil testing, stoichiometric modeling, and microbial ecology characterization to better characterize the potential use of denitrification as a two-stage process for liquefaction mitigation. / Dissertation/Thesis / Doctoral Dissertation Civil and Environmental Engineering 2016

Com uma necessidade cada vez maior da utilização de materiais mais sustentáveis na construção civil, existe uma grande busca por novas substâncias que possam suprir as nossas necessidades de maneira mais “limpa”. O presente trabalho tem como foco propor uma metodologia de dosagem para a obtenção de solos artificialmente cimentados, com determinadas resistências e rigidezes, de maneira que eles sejam duráveis, com a adição de diferentes resíduos. O solo inicialmente estudado é a areia de Osório e os resíduos são a cal de carbureto e o pó de vidro. A cal de carbureto consiste basicamente de hidróxido de cálcio e o pó de vidro de sílica em estado amorfo, gerando assim uma combinação adequada para a ocorrência das reações pozolânicas. A metodologia de dosagem proposta relaciona as propriedades dos materiais com o fator η/Biv0,28, aos sete dias de cura, sendo η a porosidade do solo artificialmente cimentado e, Biv o teor volumétrico de agente cimentante no interior do mesmo (neste caso pó de vidro e cal de carbureto). O expoente externo 0,28 se assemelha ao encontrado em outras metodologias de dosagem para solos artificialmente cimentados com cimento Portland Com a finalidade de validar a metodologia proposta, para condições distintas das analisadas, foram executados ensaios com diferentes períodos de cura (180 dias), diferentes resíduos (cinza de casca de arroz) e um solo distinto (arenito Botucatu), chegando também a correlações do comportamento mecânico destes materiais com o fator η/Biv0,28. Os dados obtidos nesta pesquisa foram comparados com os resultados de outros pesquisadores que utilizaram cal com cinza volante de usinas de carvão juntamente com cal hidratada e de outros trabalhos que utilizaram cimento Portland, todos para criar solos artificialmente cimentados, e, mais uma vez, a metodologia proposta se mostrou eficaz. Foi possível ainda normalizar em função de um valor médio todos os resultados de rigidez, resistência e durabilidade obtendo assim equações únicas em função de η/Biv0,28 para todos os materiais analisados, chegando a uma correlação com mais de 1000 pontos no caso da resistência. / With a growing need for the use of more sustainable materials in construction, there is a great search for new substances that can meet our needs in a "cleaner" way. The present work aims to propose a dosing methodology for obtaining artificially cemented soils, with certain strengths and stiffness, so that they are durable, with the addition of different residues. The initially studied soil is the Osório sand and the residues are carbide lime and glass powder. The carbide lime basically consists of calcium hydroxide and the glass powder consists of silica in the amorphous state, thus generating a great combination for the occurrence of the pozolanic reactions. The proposed dosage methodology relates the properties of the materials with the factor η/Biv0.28, at the seven days of curing, in which η is the porosity of the artificially cemented soil and Biv is the volumetric content of the cementitious agent inside it (in this case powder glass and carbide lime) The external exponent 0.28 resembles that found in other dosing methodologies for artificially cemented soils with Portland cement. The methodology led to a unique correlation for all the dosages used. To validate the hypothesis which was raised, tests were carried out with different curing periods (180 days), different residues (rice husk ash) and a different soil (Botucatu sandstone), reaching optimum correlations. The data obtained in this research were compared with the results of other researchers who used lime with fly ash from coal plants together with hydrated lime and from other works that used Portland cement, all to create artificially cemented soils, and, once again, the proposed methodology proved to be excellent. It was also possible to normalize all the results of stiffness, strength and durability thus obtaining unique equations as a function of η/Biv0.28 for all materials analyzed, reaching a correlation with more than 1000 points.

Este trabalho consiste em um estudo para viabilizar o melhoramento de um solo mole orgânico com o uso de cimentação. Foram avaliadas numerosas variáveis, como diferentes agentes cimentantes, distintos materiais juntamente com estes agentes e diferentes temperaturas de cura. Foi desenvolvida uma metodologia de processamento dessas misturas para viabilizar e otimizar a moldagem, cura, desmoldagem e ruptura dos elementos que foram analisados. Os agentes cimentantes utilizados foram a cal viva e o cimento de alta resistência inicial. Juntamente com os agentes cimentantes, foram empregados escória de aciaria, escória de forno panela, fosfogesso e gesso. As temperaturas de cura empregadas foram 20 graus Célsius e 50 graus Célsius. O solo utilizado neste estudo, uma argila mole orgânica com teor de umidade natural variando de 85% a 100%, obtido em Nova Santa Rita, Rio Grande do Sul, foi caracterizado química e fisicamente. Para caracterizá-lo quimicamente foram executados ensaios de perda ao fogo, análise elementar, difratograma de raio X e fracionamento da matéria orgânica. Na faixa de variáveis analisadas, o melhoramento deste material com o uso dos agentes cimentantes utilizados mostrou-se possível, o mais indicado para utilização é o cimento CP-V e quanto maior a dosagem adicionada maior foi a resistência obtida. Com a cal viva, observou-se um teor ótimo, a partir do qual um acréscimo de material gera uma diminuição no ganho de resistência, fato que pode em um projeto real ser prejudicial. As escórias de aciaria e de forno panela não se mostraram eficientes na avaliação do ganho de resistência quando utilizadas com os agentes cimentantes analisados neste material. O gesso e o fosfogesso proporcionaram um ganho de resistência considerável à mistura quando utilizados junto com o cimento CP-V, principalmente para altas dosagens, já com cal tiveram um efeito extremamente prejudicial, fazendo o ganho de resistência da mistura solo/cal viva cair em até 50%. A temperatura de cura mostrou-se eficaz na aceleração das reações pozolânicas tanto para a cal quanto para o cimento. Foi possível adaptar uma metodologia de dosagem em função do fator porosidade/teor volumétrico de cimento apenas para as misturas solo/cimento. / The focus of this study is enabling the improvement of an organic soft soil by means of pozzolanic reaction. A several variables were evaluated, such as type of stabilization, pure or mixed with other materials and two distinct curing temperatures. In order to facilitate and optimize the molding, curing, unmolding and rupture of the specimens, a methodology was developed for processing the mixtures. The binding agents used were cement and quicklime. Mixed with the binding agents, electric arch furnace slag, ladle furnace slag, gypsum and phosphoric gypsum were added. The used curing temperatures were 20 and 50 degrees Celsius. The soil used in this study is a very soft organic clay with natural moisture content ranging from 85% to 100% obtained in Nova Santa Rita, Rio Grande do Sul, Brazil. It was chemically and physically characterized. The chemical characterization consisted of loss of ignition tests, elemental analysis, X-ray diffraction and the fractionation of organic matter. The improvement of this material with the use of the previously mentioned binders proved possible within the range of variation of the analyzed variables. The best results were obtained when using cement, and the larger the content of cement the greater the added strength. As for the quicklime, an optimum lime content was observed. From this optimum point onwards any addition of quicklime will result in diminishment of the gained strength as such, when utilizing this method in a real project, one must be aware of this diminishment so as to avoid detrimental results. Utilized with the binding agents analyzed in this material, the electric arc furnace slag and the ladle slag were shown to be inefficient when evaluating the gain of strength. Gypsum and phosphoric gypsum led to considerable gain in strength when mixed with cement, especially in high dosages. When mixed with quicklime, however, the effect was a drop of up to 50% in the gain of strength. The curing temperature was effective in accelerating pozzolanic reactions for both the cement and the quicklime. It was possible to adapt a dosing methodology using the function porosity/volumetric cement content into soil/cement mixtures.

abstract: Microbially induced calcium carbonate precipitation (MICP) is attracting increasing attention as a sustainable means of soil improvement. While there are several possible MICP mechanisms, microbial denitrification has the potential to become one of the preferred methods for MICP because complete denitrification does not produce toxic byproducts, readily occurs under anoxic conditions, and potentially has a greater carbonate yield per mole of organic electron donor than other MICP processes. Denitrification may be preferable to ureolytic hydrolysis, the MICP process explored most extensively to date, as the byproduct of denitrification is benign nitrogen gas, while the chemical pathways involved in hydrolytic ureolysis processes produce undesirable and potentially toxic byproducts such as ammonium (NH4+). This thesis focuses on bacterial denitrification and presents preliminary results of bench-scale laboratory experiments on denitrification as a candidate calcium carbonate precipitation mechanism. The bench-scale bioreactor and column tests, conducted using the facultative anaerobic bacterium Pseudomonas denitrificans, show that calcite can be precipitated from calcium-rich pore water using denitrification. Experiments also explore the potential for reducing environmental impacts and lowering costs associated with denitrification by reducing the total dissolved solids in the reactors and columns, optimizing the chemical matrix, and addressing the loss of free calcium in the form of calcium phosphate precipitate from the pore fluid. The potential for using MICP to sequester radionuclides and metal contaminants that are migrating in groundwater is also investigated. In the sequestration process, divalent cations and radionuclides are incorporated into the calcite structure via substitution, forming low-strontium calcium carbonate minerals that resist dissolution at a level similar to that of calcite. Work by others using the bacterium Sporosarcina pasteurii has suggested that in-situ sequestration of radionuclides and metal contaminants can be achieved through MICP via hydrolytic ureolysis. MICP through bacterial denitrification seems particularly promising as a means for sequestering radionuclides and metal contaminants in anoxic environments due to the anaerobic nature of the process and the ubiquity of denitrifying bacteria in the subsurface. / Dissertation/Thesis / M.S. Engineering 2013

O jet grouting é uma das técnicas de tratamento de solos mais utilizadas em todo o mundo. Consiste em jatear uma calda de cimento a altíssima velocidade no subsolo, com ou sem a adição de água e/ou ar comprimido. Os fluidos são injetados através de minúsculos bicos posicionados na extremidade de uma composição especial de hastes, que giram à velocidade constante enquanto sobem lentamente em direção à superfície do terreno. O jato remove e mistura o solo, produzindo um corpo consolidado de formato aproximadamente cilíndrico (a coluna de jet grouting). Variando os parâmetros operacionais (pressão de bombeamento, quantidade e diâmetro dos bicos, velocidade de extração das hastes, relação água/cimento da calda) e em função do tipo de solo, podem ser obtidas colunas de diâmetros variáveis dentro de um amplo intervalo. Na busca do diâmetro desejado, a escolha dos parâmetros de tratamento é feita atualmente com base em regras empíricas nem sempre pertinentes, causando muitas vezes o fracasso do tratamento ou, ainda, tornando-o excessivamente oneroso. Esta tese de doutorado apresenta um método simplificado para a previsão do diâmetro das colunas de jet grouting (sistema monofluido). Conjugando a análise racional do fenômeno físico de interação jatosolo (Modoni et al., 2006) à facilidade de aplicação característica dos métodos empíricos, propõe-se uma ferramenta ágil para a escolha da combinação mais adequada dos parâmetros de tratamento. / Jet grouting is one of the most popular ground improvement techniques all over the world. The method is based on high-speed grouting of water-cement mixtures and/or other fluids (air, water) into the subsoil. The fluids are injected through small diameter nozzles placed on a grout pipe, which is continuously rotated at a constant rate and slowly raised towards the ground surface. The jet removes and mixes the soil, producing a cemented body of quasi-cylindrical shape (the jet grouting column). Varying the operational parameters (grout pressure, number and diameter of the nozzles, monitor lifting rate, water-cement ratio of the grout) and as a function of the soil type, columns of variable diameters can be obtained in a wide interval. Currently, in search of the required diameter, the choice of the operational parameters is done on the basis of empirical rules not always relevant, causing very often the failure of the treatment or making it excessively onerous. This thesis presents a simplificated method for the prediction of the column diameter (single fluid system). Conjugating the rational analysis of the physical phenomenon of jetsoil interaction (Modoni et al., 2006) to the typical easiness of empirical methods application, an agile tool is proposed for the choice of the most appropriate combination of the operational parameters.

Este trabalho consiste em um estudo para viabilizar o melhoramento de um solo mole orgânico com o uso de cimentação. Foram avaliadas numerosas variáveis, como diferentes agentes cimentantes, distintos materiais juntamente com estes agentes e diferentes temperaturas de cura. Foi desenvolvida uma metodologia de processamento dessas misturas para viabilizar e otimizar a moldagem, cura, desmoldagem e ruptura dos elementos que foram analisados. Os agentes cimentantes utilizados foram a cal viva e o cimento de alta resistência inicial. Juntamente com os agentes cimentantes, foram empregados escória de aciaria, escória de forno panela, fosfogesso e gesso. As temperaturas de cura empregadas foram 20 graus Célsius e 50 graus Célsius. O solo utilizado neste estudo, uma argila mole orgânica com teor de umidade natural variando de 85% a 100%, obtido em Nova Santa Rita, Rio Grande do Sul, foi caracterizado química e fisicamente. Para caracterizá-lo quimicamente foram executados ensaios de perda ao fogo, análise elementar, difratograma de raio X e fracionamento da matéria orgânica. Na faixa de variáveis analisadas, o melhoramento deste material com o uso dos agentes cimentantes utilizados mostrou-se possível, o mais indicado para utilização é o cimento CP-V e quanto maior a dosagem adicionada maior foi a resistência obtida. Com a cal viva, observou-se um teor ótimo, a partir do qual um acréscimo de material gera uma diminuição no ganho de resistência, fato que pode em um projeto real ser prejudicial. As escórias de aciaria e de forno panela não se mostraram eficientes na avaliação do ganho de resistência quando utilizadas com os agentes cimentantes analisados neste material. O gesso e o fosfogesso proporcionaram um ganho de resistência considerável à mistura quando utilizados junto com o cimento CP-V, principalmente para altas dosagens, já com cal tiveram um efeito extremamente prejudicial, fazendo o ganho de resistência da mistura solo/cal viva cair em até 50%. A temperatura de cura mostrou-se eficaz na aceleração das reações pozolânicas tanto para a cal quanto para o cimento. Foi possível adaptar uma metodologia de dosagem em função do fator porosidade/teor volumétrico de cimento apenas para as misturas solo/cimento. / The focus of this study is enabling the improvement of an organic soft soil by means of pozzolanic reaction. A several variables were evaluated, such as type of stabilization, pure or mixed with other materials and two distinct curing temperatures. In order to facilitate and optimize the molding, curing, unmolding and rupture of the specimens, a methodology was developed for processing the mixtures. The binding agents used were cement and quicklime. Mixed with the binding agents, electric arch furnace slag, ladle furnace slag, gypsum and phosphoric gypsum were added. The used curing temperatures were 20 and 50 degrees Celsius. The soil used in this study is a very soft organic clay with natural moisture content ranging from 85% to 100% obtained in Nova Santa Rita, Rio Grande do Sul, Brazil. It was chemically and physically characterized. The chemical characterization consisted of loss of ignition tests, elemental analysis, X-ray diffraction and the fractionation of organic matter. The improvement of this material with the use of the previously mentioned binders proved possible within the range of variation of the analyzed variables. The best results were obtained when using cement, and the larger the content of cement the greater the added strength. As for the quicklime, an optimum lime content was observed. From this optimum point onwards any addition of quicklime will result in diminishment of the gained strength as such, when utilizing this method in a real project, one must be aware of this diminishment so as to avoid detrimental results. Utilized with the binding agents analyzed in this material, the electric arc furnace slag and the ladle slag were shown to be inefficient when evaluating the gain of strength. Gypsum and phosphoric gypsum led to considerable gain in strength when mixed with cement, especially in high dosages. When mixed with quicklime, however, the effect was a drop of up to 50% in the gain of strength. The curing temperature was effective in accelerating pozzolanic reactions for both the cement and the quicklime. It was possible to adapt a dosing methodology using the function porosity/volumetric cement content into soil/cement mixtures.

Este trabalho consiste em um estudo para viabilizar o melhoramento de um solo mole orgânico com o uso de cimentação. Foram avaliadas numerosas variáveis, como diferentes agentes cimentantes, distintos materiais juntamente com estes agentes e diferentes temperaturas de cura. Foi desenvolvida uma metodologia de processamento dessas misturas para viabilizar e otimizar a moldagem, cura, desmoldagem e ruptura dos elementos que foram analisados. Os agentes cimentantes utilizados foram a cal viva e o cimento de alta resistência inicial. Juntamente com os agentes cimentantes, foram empregados escória de aciaria, escória de forno panela, fosfogesso e gesso. As temperaturas de cura empregadas foram 20 graus Célsius e 50 graus Célsius. O solo utilizado neste estudo, uma argila mole orgânica com teor de umidade natural variando de 85% a 100%, obtido em Nova Santa Rita, Rio Grande do Sul, foi caracterizado química e fisicamente. Para caracterizá-lo quimicamente foram executados ensaios de perda ao fogo, análise elementar, difratograma de raio X e fracionamento da matéria orgânica. Na faixa de variáveis analisadas, o melhoramento deste material com o uso dos agentes cimentantes utilizados mostrou-se possível, o mais indicado para utilização é o cimento CP-V e quanto maior a dosagem adicionada maior foi a resistência obtida. Com a cal viva, observou-se um teor ótimo, a partir do qual um acréscimo de material gera uma diminuição no ganho de resistência, fato que pode em um projeto real ser prejudicial. As escórias de aciaria e de forno panela não se mostraram eficientes na avaliação do ganho de resistência quando utilizadas com os agentes cimentantes analisados neste material. O gesso e o fosfogesso proporcionaram um ganho de resistência considerável à mistura quando utilizados junto com o cimento CP-V, principalmente para altas dosagens, já com cal tiveram um efeito extremamente prejudicial, fazendo o ganho de resistência da mistura solo/cal viva cair em até 50%. A temperatura de cura mostrou-se eficaz na aceleração das reações pozolânicas tanto para a cal quanto para o cimento. Foi possível adaptar uma metodologia de dosagem em função do fator porosidade/teor volumétrico de cimento apenas para as misturas solo/cimento. / The focus of this study is enabling the improvement of an organic soft soil by means of pozzolanic reaction. A several variables were evaluated, such as type of stabilization, pure or mixed with other materials and two distinct curing temperatures. In order to facilitate and optimize the molding, curing, unmolding and rupture of the specimens, a methodology was developed for processing the mixtures. The binding agents used were cement and quicklime. Mixed with the binding agents, electric arch furnace slag, ladle furnace slag, gypsum and phosphoric gypsum were added. The used curing temperatures were 20 and 50 degrees Celsius. The soil used in this study is a very soft organic clay with natural moisture content ranging from 85% to 100% obtained in Nova Santa Rita, Rio Grande do Sul, Brazil. It was chemically and physically characterized. The chemical characterization consisted of loss of ignition tests, elemental analysis, X-ray diffraction and the fractionation of organic matter. The improvement of this material with the use of the previously mentioned binders proved possible within the range of variation of the analyzed variables. The best results were obtained when using cement, and the larger the content of cement the greater the added strength. As for the quicklime, an optimum lime content was observed. From this optimum point onwards any addition of quicklime will result in diminishment of the gained strength as such, when utilizing this method in a real project, one must be aware of this diminishment so as to avoid detrimental results. Utilized with the binding agents analyzed in this material, the electric arc furnace slag and the ladle slag were shown to be inefficient when evaluating the gain of strength. Gypsum and phosphoric gypsum led to considerable gain in strength when mixed with cement, especially in high dosages. When mixed with quicklime, however, the effect was a drop of up to 50% in the gain of strength. The curing temperature was effective in accelerating pozzolanic reactions for both the cement and the quicklime. It was possible to adapt a dosing methodology using the function porosity/volumetric cement content into soil/cement mixtures.

The 1999 Kocaeli Earthquake (M=7.4) struck northwestern Turkey on August 17, 1999 and caused significant damage in urban areas located along Izmit Bay. The sites that suffered the greatest damages were located primarily in areas of poorest soil conditions, typically containing soft clays and silts and/or loose, liquefiable sands. Because the affected region is heavily developed with infrastructure and there is a preponderance of poor soils, a wide range of soil improvement measures had been used to mitigate anticipated earthquake damages throughout the region. Following the earthquake and significant aftershocks, Virginia Tech researchers traveled to Turkey to investigate the affected area to document geotechnical field performance. Primary focus of the Virginia tech team was given to investigating the performance of improved soil sites and reinforced soil structures. The sites were subjected to ground motions ranging from about 0.10g to 0.35g. The site locations ranged from 0 to 35 km from the zone of energy release. This dissertation presents in detail, the findings from the two most instructive sites. The investigation of these sties involved field reconnaissance, field and laboratory testing of soils, seismic analysis, numerical modeling, and other analytical work. / Ph. D.

Soil improvement methods are used to minimize the consequences of liquefaction by changing the characteristics and/or response of a liquefiable soil deposit. When considering sites with previous development, the options for soil improvement are limited. Traditional methods, such as compaction and vibratory techniques, are difficult to employ because of adverse effects on adjacent structures. One potential method for soil improvement against soil liquefaction in developed sites is accelerated drainage through in situ vertical drains. Vertical drains expedite the dissipation of excess pore water pressures by reducing the length of the pore water drainage path. For more than thirty years, vertical gravel drains or stone columns have been employed to ensure the excess pore water pressure ratio remains below a prescribed maximum value. In recent years, the use of prefabricated vertical drains (PVDs) has increased because the drains can be installed with less site disruption than with traditional soil improvement methods. To date, little-to-no field or experimental verification is available regarding the seismic performance of sites treated with PVDs. The effectiveness of PVDs for liquefaction remediation was evaluated via small-scale centrifuge testing and full-scale field testing. A small-scale centrifuge test was performed on an untreated soil deposit and on a soil deposit treated with small-scale vertical drains. Compared to the untreated condition, the presence of the small-scale vertical drains provided numerous benefits including smaller magnitudes of excess pore water pressure generation and buildup, smaller induced cyclic shear strains, reduced times for pore pressure dissipation, and smaller permanent horizontal and vertical displacements. In addition, full-scale in situ field experiments were performed in an untreated soil deposit and in a soil deposit treated with full-scale PVDs using a vibrating mandrel as the dynamic source. In the untreated test area, the maximum induced excess pore pressure ratio reached about 0.95. In the treated test area, the vibratory installation of the first few drains generated significant excess pore pressures; however, significant excess pore pressures were not generated during the vibratory installation of additional drains because of the presence of the adjacent drains. Additionally, the vibratory installation of the drains caused significant settlement and significantly altered the shear wave velocity of the sand. Dynamic shaking after installation of all of the drains induced small accelerations, small cyclic shear strains, and negligible excess pore water pressures in the soil. The results of the field experiment indicate that the prefabricated vertical drains were effective at dissipating excess pore water pressures during shaking and densifying the site. / text

A fundação direta é limitada em solos colapsíveis pois o efeito do colapso é mais acentuado nas camadas superficiais do terreno. Diante deste cenário, nesta pesquisa foi avaliada uma nova técnica de reforço de solo com a finalidade de viabilizar o uso de fundações diretas para controlar e reduzir o recalque das fundações devido ao colapso. A técnica proposta consiste na execução de colunas de solo laterítico compactado, inseridas na camada colapsível. O processo de compactação das colunas provoca a densificação do solo ao seu entorno, e consequentemente reduz a compressibilidade da massa de solo reforçado. Neste trabalho também foi verificado se a compactação radial do solo em torno da coluna estava relacionada com a geometria do pilão utilizado. Para compreender estes efeitos, foram realizados ensaios em laboratório, em escala reduzida, para avaliar qualitativamente duas geometrias de pilões diferentes, uma cônica e outra cilíndrica (tradicionalmente utilizada na engenharia de fundações). Como resultado desta análise, concluiu-se que a geometria cônica apresentou uma maior compactação lateral do solo no entorno da coluna, comparando-se com a geometria cilíndrica. Para analisar a eficiência do método de reforço proposto, foram executadas e ensaiadas colunas em escala real no solo colapsível do Campo Experimental de Fundações da Escola de Engenharia de São Carlos da Universidade de São Paulo. Para esta análise, foram realizadas provas de carga em placa em solo com e sem reforço. Os ensaios foram realizados com e sem a inundação do terreno para se verificar o efeito da inundação no comportamento do solo reforçado. Desta forma, foi possível avaliar o ganho de capacidade de carga e a redução dos recalques devido à execução dos elementos de reforço ensaiados. Por meio dos resultados obtidos nas provas de carga em placa constatou-se a eficiência do método na redução do recalque e no aumento da capacidade de carga. Comparando-se as curvas dos ensaios de placa com e sem reforço, observou-se um aumento da capacidade de carga de quatro vezes em consequência da instalação do reforço. / The use of shallow foundation is limited on collapsible soils due to the collapse effect pronounced in superficial layers. Considering this scenario, this research presents a new soil reinforcement technique in order to reduce the settlement of foundations due to the collapse. The proposed technique involves the execution of compacted lateritic soil columns inserted into the collapsible soil. The compaction of these columns causes the densification of the surrounding soil and consequently reduces compressibility of the reinforced soil mass. In this work, it was also evaluated the effect of the hummer geometry on the radial compaction of the soil around the column. For this evaluation, laboratory tests were conducted using two different geometries of hammer, one conical and other cylindrical (traditionally used in foundation engineering). The results showed that the conical geometry provided greater lateral compaction of the soil around the column compared to the cylindrical geometry. The efficacy of the compacted lateritic soil columns was evaluated by the results of plate load tests performed on the collapsible soil of the Experimental Field of the University of São Paulo in São Carlos city. The tests were carried out on a plate installed in the ground with and without reinforcement. The gain on bearing capacity and settlement reductions due to the reinforcement were verified by using these tests results. The tests were performed in natural and in flooded condition to verify the efficiency of the soil reinforcement in the flooding situation. Based on the results of the plate load tests, it was verified the efficacy of the method to reduce the effect of collapse. The load-displacement curves of the load tests showed that the solution evaluated provided an increase of 400% on the ultimate capacity of the plate installed on the collapsible soil tested in this investigation.