Made available in DSpace on 2016-12-08T17:19:38Z (GMT). No. of bitstreams: 1 Paginas Iniciais.pdf: 93423 bytes, checksum: 5bfa2c63131c8e51c48b7c76006ab53a (MD5) Previous issue date: 2011-02-28 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / During the execution of this work it was developed the Project of an equipment to perform the trial of torsion following standard rules available for this test. This equipment was built and tested in the last two years and the main focus of this development was to obtain a low-cost equipment with capacity to run trials with good precision. In addition to this equipment it was built accessories to perform the measurement of torsion and angular variation between two points. The torquemeter was designed in aluminum and using straingage with a circuit in full bridge to the measurement of the torque. The project of this torquemeter followed the same standards as the ones available in the market; however the cost of its building was also very low. The other accessory built was the torsiometer and it wasn´t found in literature (neither commercially) any similar project to the one developed in this work. The measurement of the angular variation between two points was also performed using strain-gage with circuit in ¼ bridge. The curve of torque x torsion angle is the result of the torsion trial from where two main data are obtained: the material shear module and the rupture module (the amount of the shear tension for the material rupture). To test the set of equipment and accessories trials were performed with four different materials: 1035 steel, 1020 steel, aluminum and brass. These results were compared with the ones obtained in the traction trial and with data available in the literature. Along this work it is shown clearly the building details, the costs, the mathematical equation of elastic-plastic torsion and the quality of the results obtained with the equipment developed and built during the execution of this dissertation. / Durante o cumprimento deste trabalho foi criado o projeto de um equipamento para efetuar o ensaio de torção seguindo padrões de normas disponíveis para o teste. O equipamento foi construído e testado nos últimos dois anos, e o principal objetivo de tal desenvolvimento foi obter um aparelho de baixo custo com capacidade de realizar ensaios com boa precisão. Além do equipamento, foram construídos acessórios para as medições de torque e da variação angular entre dois pontos. Projetou-se e construiu-se o torquímetro em alumínio empregando extensometria com circuito em ponte completa para a medição do torque. O projeto do torquímetro atingiu os mesmos padrões daqueles que estão disponíveis no mercado, porém com custo muito baixo. Outro acessório construído foi o torsiômetro, e não se encontrou na literatura (ou comercialmente) nenhum projeto similar ao desta dissertação. A medição da variação angular também se fez com extensometria com circuito em ¼ de ponte. A curva do torque × ângulo de torção é o resultado do ensaio de torção do qual são alcançadas duas propriedades: o módulo de elasticidade transversal do material e o módulo de ruptura (o valor da tensão cisalhante para a ruptura do material). A fim de testar o conjunto equipamento e acessórios, foram realizados ensaios com quatro materiais diferentes: aço 1035, aço 1020, alumínio e latão. Compararam-se os resultados com aqueles atingidos no ensaio de tração e com os dados achados na literatura. Ao longo deste trabalho mostram-se com clareza os detalhes construtivos, os custos, o equacionamento matemático da torção elastoplástica e a precisão dos resultados obtidos com o equipamento desenvolvido com esta dissertação.

Objetivo: verificar, através de análise por elementos finitos e de uma etapa experimental, se o ensaio de torção, como ensaio de resistência de interfaces adesivas, seria capaz de levar os espécimes à ruptura sob tensões de cisalhamento e se ele seria vantajoso em relação ao ensaio de cisalhamento. Material e Método: para a análise de tensões pelo método dos elementos finitos, os modelos tridimensionais dos dois ensaios foram constituídos por cilindro de compósito, uma camada de adesivo e um cilindro de dentina com maior diâmetro. Os espécimes foram simulados nos tamanhos convencional e micro, numa proporção de 5:1, exceto pela camada de adesivo, de espessura constante. Foram simulados dois módulos de elasticidade do cilindro de compósito (híbrido e baixa viscosidade) e, para os modelos de cisalhamento, variaram as distâncias entre o local de carregamento e a interface adesiva. Foram analisados os valores das tensões máxima principal e cisalhamento máximo e a proporção entre elas ao longo da interface dentina/adesivo; a direção dos vetores de tensão máxima principal; e o local dos picos de tensão. Na etapa experimental, foram construídos cilindros de compósito (híbrido e de baixa viscosidade) nos tamanhos convencional e micro, aderidos a uma superfície de dentina bovina por meio de três sistemas adesivos. O esforço de cisalhamento foi aplicado com fio ortodôntico a uma distância da interface adesiva baseada na literatura e, para os ensaios de torção e microtorção, foram desenvolvidos dispositivos específicos para a aplicação do torque. Os resultados de tensão de ruptura foram submetidos à análise de variância e teste de Tukey. Todas as superfícies de fratura foram analisadas por meio de microscopia eletrônica de varredura. As frequências das fraturas adesivas e coesivas foram submetidas ao teste exato de Fisher e relacionadas aos valores de tensão de ruptura experimentais, bem como suas características foram relacionadas às tensões observadas pelo método dos elementos finitos. Os parâmetros experimentais de comparação entre os ensaios foram: poder de detecção de diferenças significantes entre os sistemas adesivos; constância da classificação relativa dos mesmos ao alterar as configurações dos ensaios e frequências dos tipos de fratura. Resultados: ambos os ensaios apresentaram um estado complexo de tensões com distribuição nãouniforme. Não foi notada menor sensibilidade dos resultados do ensaio de torção às mudanças das configurações do teste, nem uma maior capacidade para detectar diferenças significantes entre os adesivos do que o ensaio de cisalhamento. O ensaio de torção apresentou maior dificuldade de execução prática e maiores frequências de fraturas coesivas, porém, estas ocorreram mais no cilindro de compósito, enquanto no cisalhamento, ocorreram mais na dentina. Não houve certeza sobre o tipo de tensão que tenha levado à fratura nos dois tipos de ensaio. Conclusões: embora não se tenha certeza se o ensaio de torção leva à ruptura sob tensões de cisalhamento e tenham sido notadas desvantagens práticas de execução do ensaio, este apresenta uma característica que sugere novos estudos: ocorrência de fraturas coesivas principalmente no cilindro de compósito, o que poderia, talvez, ser resolvido com o reforço do material do cilindro. / Objective: verify, by finite element analysis (FEA) and laboratorial tests, if the torsion bond strength test is able to lead adhesive interface to fracture under shear stress and if it would be more advantageous than wire-loop shear test. Material and method: for stress analysis by finite element method, the 3D models of both tests consisted of a resin composite cylinder, an adhesive layer and a dentin cylinder with a larger diameter. The specimens were simulated with both conventional and micro sizes, in a 5:1 ratio, except for the adhesive layer, kept in a constant thickness. Two values of elastic modulus were simulated for the composite cylinder (hybrid and flowable). For the shear models, different distances were simulated between the load point and the adhesive interface. The values of maximum principal and maximum shear stresses and the ratio between both types were analyzed along the adhesive interface, as well as the maximum principal vectors direction and the local of stress peaks. For the laboratorial tests, cylinders of hybrid and flowable resin composite, in both micro and conventional sizes, were adhered onto dentin surfaces by means of 3 adhesive systems. The shear load was applied with a metallic wire-loop at distances from the interface based on literature and, for the torsion test, the torque was applied by specific apparatus developed for that. The values of bond strength were submitted to analysis of variance and Tukeys test. Fracture surfaces were analyzed by scan electron microscopy. The frequencies of fracture types were compared by Fishers test and its relationship with bond strength values were verified, as well as the relationship between fracture features and stresses results obtained by FEA. The experimental parameters for comparison between both tests were: ability for detecting significant differences among the 3 adhesives, change of adhesives ranking while varying the test configuration and frequency of fracture types. Results: both tests showed non-uniform stress distribution. It was not noticed a lower sensibility of torsion test results to changes of the test configurations, nor a major ability for detecting significant differences among the adhesive systems than showed the shear test. The torsion test showed to be more difficult to execute and exhibited a higher frequency of cohesive fracture, however, that occurred most of the time in the composite cylinder, while it occurred mainly in dentin in shear test. There was no certainty about the kind of stress that lead to specimen fracture in both tests. Conclusions: although there was no certainty about the kind of stress that lead to specimen fracture and the major difficulty of execution of torsion test, this test exhibited a feature that suggests future studies: the occurrence of cohesive fracture mainly in the composite cylinder, which could be solved, maybe, by reinforcing the cylinder material.

Objetivo: verificar, através de análise por elementos finitos e de uma etapa experimental, se o ensaio de torção, como ensaio de resistência de interfaces adesivas, seria capaz de levar os espécimes à ruptura sob tensões de cisalhamento e se ele seria vantajoso em relação ao ensaio de cisalhamento. Material e Método: para a análise de tensões pelo método dos elementos finitos, os modelos tridimensionais dos dois ensaios foram constituídos por cilindro de compósito, uma camada de adesivo e um cilindro de dentina com maior diâmetro. Os espécimes foram simulados nos tamanhos convencional e micro, numa proporção de 5:1, exceto pela camada de adesivo, de espessura constante. Foram simulados dois módulos de elasticidade do cilindro de compósito (híbrido e baixa viscosidade) e, para os modelos de cisalhamento, variaram as distâncias entre o local de carregamento e a interface adesiva. Foram analisados os valores das tensões máxima principal e cisalhamento máximo e a proporção entre elas ao longo da interface dentina/adesivo; a direção dos vetores de tensão máxima principal; e o local dos picos de tensão. Na etapa experimental, foram construídos cilindros de compósito (híbrido e de baixa viscosidade) nos tamanhos convencional e micro, aderidos a uma superfície de dentina bovina por meio de três sistemas adesivos. O esforço de cisalhamento foi aplicado com fio ortodôntico a uma distância da interface adesiva baseada na literatura e, para os ensaios de torção e microtorção, foram desenvolvidos dispositivos específicos para a aplicação do torque. Os resultados de tensão de ruptura foram submetidos à análise de variância e teste de Tukey. Todas as superfícies de fratura foram analisadas por meio de microscopia eletrônica de varredura. As frequências das fraturas adesivas e coesivas foram submetidas ao teste exato de Fisher e relacionadas aos valores de tensão de ruptura experimentais, bem como suas características foram relacionadas às tensões observadas pelo método dos elementos finitos. Os parâmetros experimentais de comparação entre os ensaios foram: poder de detecção de diferenças significantes entre os sistemas adesivos; constância da classificação relativa dos mesmos ao alterar as configurações dos ensaios e frequências dos tipos de fratura. Resultados: ambos os ensaios apresentaram um estado complexo de tensões com distribuição nãouniforme. Não foi notada menor sensibilidade dos resultados do ensaio de torção às mudanças das configurações do teste, nem uma maior capacidade para detectar diferenças significantes entre os adesivos do que o ensaio de cisalhamento. O ensaio de torção apresentou maior dificuldade de execução prática e maiores frequências de fraturas coesivas, porém, estas ocorreram mais no cilindro de compósito, enquanto no cisalhamento, ocorreram mais na dentina. Não houve certeza sobre o tipo de tensão que tenha levado à fratura nos dois tipos de ensaio. Conclusões: embora não se tenha certeza se o ensaio de torção leva à ruptura sob tensões de cisalhamento e tenham sido notadas desvantagens práticas de execução do ensaio, este apresenta uma característica que sugere novos estudos: ocorrência de fraturas coesivas principalmente no cilindro de compósito, o que poderia, talvez, ser resolvido com o reforço do material do cilindro. / Objective: verify, by finite element analysis (FEA) and laboratorial tests, if the torsion bond strength test is able to lead adhesive interface to fracture under shear stress and if it would be more advantageous than wire-loop shear test. Material and method: for stress analysis by finite element method, the 3D models of both tests consisted of a resin composite cylinder, an adhesive layer and a dentin cylinder with a larger diameter. The specimens were simulated with both conventional and micro sizes, in a 5:1 ratio, except for the adhesive layer, kept in a constant thickness. Two values of elastic modulus were simulated for the composite cylinder (hybrid and flowable). For the shear models, different distances were simulated between the load point and the adhesive interface. The values of maximum principal and maximum shear stresses and the ratio between both types were analyzed along the adhesive interface, as well as the maximum principal vectors direction and the local of stress peaks. For the laboratorial tests, cylinders of hybrid and flowable resin composite, in both micro and conventional sizes, were adhered onto dentin surfaces by means of 3 adhesive systems. The shear load was applied with a metallic wire-loop at distances from the interface based on literature and, for the torsion test, the torque was applied by specific apparatus developed for that. The values of bond strength were submitted to analysis of variance and Tukeys test. Fracture surfaces were analyzed by scan electron microscopy. The frequencies of fracture types were compared by Fishers test and its relationship with bond strength values were verified, as well as the relationship between fracture features and stresses results obtained by FEA. The experimental parameters for comparison between both tests were: ability for detecting significant differences among the 3 adhesives, change of adhesives ranking while varying the test configuration and frequency of fracture types. Results: both tests showed non-uniform stress distribution. It was not noticed a lower sensibility of torsion test results to changes of the test configurations, nor a major ability for detecting significant differences among the adhesive systems than showed the shear test. The torsion test showed to be more difficult to execute and exhibited a higher frequency of cohesive fracture, however, that occurred most of the time in the composite cylinder, while it occurred mainly in dentin in shear test. There was no certainty about the kind of stress that lead to specimen fracture in both tests. Conclusions: although there was no certainty about the kind of stress that lead to specimen fracture and the major difficulty of execution of torsion test, this test exhibited a feature that suggests future studies: the occurrence of cohesive fracture mainly in the composite cylinder, which could be solved, maybe, by reinforcing the cylinder material.

Skador och slitage på brokonstruktioner av betong beror på tösalter, armeringskorrosion och frostcykler i kombination med ökande trafiklaster. Den i särklass vanligaste reparationsåtgärden för dessa typer av skador är att göra en pågjutning. Metoden går ut på att först avlägsna det skadade eller dåliga betongskiktet innan en ny betong pågjuts. För att bedöma hur bra ett pågjutningsarbete är behöver vidhäftningen mellan den nya pågjutningen och gamla betongen fastställas. Den mest använda metoden för detta syfte är s.k. dragprovning som bestämmer draghållfastheten i vidhäftningen. I praktiken är dock vidhäftningens skjuvhållfasthet av större intresse och approximeras därför utifrån draghållfastheten, men med en vridprovning kan fogens skjuvhållfasthet direkt bestämmas. Syftet med examensarbetet är att undersöka vridprovningens reliabilitet för att i framtiden eventuellt kunna bestämma skjuvhållfastheten på ett mer direkt och korrekt sätt. I examensarbetet undersöks vridprovningens reliabilitet för att bestämma skjuvhållfastheten i fogen mellan pågjutning och gammal betong med dragprovningen som referens. Studien omfattar åtta parallellprovningar av drag- och vridprovningar där den lägst uppmätta skjuvhållfastheten jämförs med fogens dimensioneringsvärde enligt EK2. Examensarbetet görs i samband med Spårväg city-projektet vid Sergels torg, Stockholm, i uppdrag från Trafikkontoret i Stockholm och CBI Betonginstitutet. Resultaten från provningarna visade att fogbrott, som är av intresse, var vanligare för dragprovning än vridprovning. Den stora spridningen av mätvärdena var en konsekvens av det låga antalet fogbrott för vridprovningen. Medelvärdet för fogens draghållfasthet på 1,43 MPa tyder på god vidhäftning, men fogens skjuvhållfasthet bestämd med vridprovning visar ett medelvärde på endast 1,61 MPa och understiger det förväntade värdet på ca dubbla draghållfastheten d.v.s. 2,9 MPa. Intressant är att det lägst uppmätta värdet på skjuvhållfastheten på 0,83 MPa var ovanligt lågt, men fortfarande större än dimensioneringsvärdet enligt EK2 på 0,59 MPa. Studien har visat att vridprovningen är en svårtydlig metod för att bestämma skjuvhållfastheten i fogen mellan pågjutning och gammal betong. Anledningen till detta beror främst av tre faktorer, för få antal provningar, den stora spridningen av mätvärden samt de svårtolkade brottmoderna från vridprovningen. Trots detta tros vridprovningen vara en framtida metod för konstruktörer och beställare eftersom metoden är ett bra mätverktyg för att bestämma skjuvhållfastheten för pågjutna betongkonstruktioner som helhet och inte bara i fogen mellan pågjutning och befintlig betong. / Damages and wear on concrete bridges are due to de-icing salt or salt water, reinforcement corrosion and repeated freeze-thaw cycles in combination with increasing traffic loads. The most common repair operation for these types of damages is to remove the deteriorated concrete and replace it with a new concrete overlay. To evaluate how well a bonded concrete overlay is, the bond strength between the new and old concrete has to be determined. The most widely used method for this purpose is the so-called pull off test to determine the tensile bond strength. In practice however the shear bond strength is of greater interest and is therefore approximated on the basis of tensile bond strength, but with a torsion test the shear bond strength can be directly determined. The purpose of this study is to investigate the reliability of the torsion test to determine the shear bond strength in a more direct and accurate manner. The thesis examines the reliability of the torsion test to determine the shear bond strength with the pull off test as a reference. The study includes eight parallel tests of pull off and torsion tests where the lowest measured shear bond strength is compared with the calculated design value of shear bond strength according to EC2. The work is done in connection with the Spårväg city project at Sergels torg, Stockholm, in collaboration with the Traffic Administration Office in Stockholm and CBI, the Swedish Cement and Concrete Research Institute. The results of the tests showed that the failures in the interface between new and old concrete, which are of interest, were more common for pull off tests than torsion tests where only two of the eight test samples showed failure in the interface. The wide scatter of the measured values is a consequence of the low number of failures in the interface. The mean value of the tensile bond strength was 1,43 MPa, which indicates good bond strength. The shear bond strength however made with torsion tests show a mean value of only 1,61 MPa. Lower than the expected value of about twice the tensile bond strength of 2,86 MPa. Interestingly, the lowest measured value of the shear bond strength of 0,83 MPa was unusually low, but still higher than the calculated design value of shear bond strength of 0.59 MPa according to EC2. The study has shown that torsion test is a difficult method for determining the shear strength of the bond between the new and old concrete. The reason for this is mainly due to three factors, the low number of tests, the large scatter of values, and the difficulty to interpret failures of the test samples. Despite this the torsion test seems to be a future method for structural engineers and contractors as a tool to determine shear strength for repaired concrete structures in general and not only the shear bond strength.

L’excellente coulabilité, les coûts de production relativement bas, et ratio poids/résistance mécanique élevé des alliages de fonderie Al-Si-Mg en font une des solutions les plus intéressantes dans le secteur automobile ainsi que dans le domaine aérospatial. Toutefois, il est bien connu que la durée de vie de ces composants moulés à grand nombre de cycles (105 < Nf < 107 cycles) est sévèrement réduite lorsque des défauts de fonderie (notamment pores et oxydes) sont débouchants et/ou subsurfaciques sont présents. Ces défauts concentrent les contraintes et peuvent considérablement réduire la période d’amorçage des fissures de fatigue en fonction de leur taille, forme et des caractéristiques microstructurales du matériau. Les défauts internes (à partir desquels les fissures peuvent amorcer et propager sans interaction avec l’air ambiant) ainsi que les défauts de surface (ceux qui sont placés à la surface et en contact direct avec l’air ambiant) vont également nuire la durée de vie des composants moulés. Toutefois, dans le cas des défauts internes, les coefficients de sécurité préconisés par les règles de conception ne font pas intervenir la distance de défaut par rapport à la surface. Le suivi de fissures de fatigue effectué à la surface d’éprouvettes macroscopiques de traction indique que la présence d’un défaut avec une taille supérieure à celle des fissures microstructuralement courtes (√A ≈ 500 μm, taille contrôlée par la SDAS) produit une remarquable réduction de la durée vie. En revanche, la durée de vie n’est pas affectée lorsqu’un défaut plus petit (√A ≈ 300 μm) est présent à la surface car l’amorçage et les premiers stades de propagation sont encore influencés par la SDAS. Les essais de fatigue en torsion pure montrent que la morphologie des surfaces de rupture est fortement influencée par le niveau de contrainte. De plus, le nombre de cycles à l’amorçage est réduit par rapport à la traction. Cet amorçage est multi-site et plusieurs fissures peuvent croitre simultanément au cours de la durée de vie d’une éprouvette, la rupture finale se produisant lors de la jonction de certaines de ces fissures. La propagation des fissures en torsion est largement influencée par la cristallographie locale et les retassures ne semblent pas être des sites de nucléation préférentiels. Les durées de vie odes échantillons macroscopiques contenant défauts artificiel internes (Øeq ≈ 2 mm) sont pratiquement similaires à celles obtenues avec un matériau de référence. L’amorçage et la propagation de fissures internes a été rarement observé lors des expériences de tomographie synchrotron. Dans les rares cas où de telles fissures ont pu être observées, le chemin de fissuration semble fortement influencé par la cristallographie alors que les fissures amorcées depuis la surface se propagent globalement en mode I. La vitesse de propagation des fissures internes est très inférieure à celle des fissures se propageant à partir de la surface. / The excellent castability, relatively low production costs, and high strength to weight ratios make Al-Si-Mg cast alloys an attractive choice for use in cheaper and lighter engineering components, in both automotive and aerospace industries. However, it is well known that High Cycle Fatigue (HCF) lives (105 < Nf < 107 cycles) of cast components are severely reduced when casting defects (notably pores and oxides) are present at the free surface or subsurface. They act as stress raisers which can considerably reduce the crack incubation period depending on their size, shape and the microstructural features of the surrounding material. Internal casting defects are of special interest to this work. The application of safety coefficients considers that all casting defects present in a component have the same deleterious effect and no attention is paid, for example, to their distance to the free surface. In other words, internal defects (corresponding to the case where the depth of the defect allows crack nucleation and propagation to essentially occur without interaction with the air environment) are considered as damaging to fatigue life as surface defects (those placed at the free surface and in contact with the air environment). Surface crack monitoring performed on uniaxial fatigue specimens indicates that the presence of a surface microshrinkage exceeding the size of microstructurally small cracks (√A ≈ 500 μm, controlled by the SDAS) readily nucleates a fatigue cracks producing steady crack propagation and remarkable reduction in the expected fatigue life. A smaller surface defect (√A ≈ 300 μm) nucleated a crack that did not reduced the expected fatigue life as in this case early stages of propagation are still nfluenced by the SDAS. Pure torsional cycling reveals that the morphology of fracture surfaces is highly influenced by the stress level. In general, torsional fatigue behaviour is described by having reduced (with respect to uniaxial testing) and multisite crack nucleation periods. Several dominant cracks can evolve simultaneously and the final failure occurs by the linkage of some of those cracks. Crack propagation is controlled by the crystallography and pores do not appear to be preferential nucleation sites. S-N curves show that macroscopic specimens containing Øeq ≈ 2 mm internal artificial defect produce similar fatigue lives to those obtained with a defect-free material. Internal crack nucleation was rarely observed during synchrotron tomography experiments; instead the fatal cracks initiated from much smaller surface defects. Tomographic images show that, in the case of internal propagation, crystallographic paths are formed while surface cracks propagate in mode I. The crack growth rate of internal cracks is much smaller than that of cracks propagating from the free surface.