Estudaram-se os efeitos da fração volumétrica de carboneto eutético e da dureza da matriz sobre a resistência à fadiga térmica de ferros fundidos brancos multicomponentes. Utilizaram-se ligas do sistema Fe-4Cr-V-2Mo-2W-2C, V variando entre 5 e 8% e Fe-4Cr-8V-Mo-2W-2C, Mo variando entre 2 e 5%. Corpos-de-prova com a geometria de discos com seção variável foram temperados e revenidos para obtenção de dois valores de microdureza da matriz: 450 HV e 650 HV. Conduziram-se ensaios de fadiga térmica por 100 e 500 ciclos. Cada ciclo envolveu aquecimento por indução da superfície até a temperatura de 600°C em 10 segundos e subseqüente resfriamento em água por 45 segundos (equalização das temperaturas do núcleo e da superfície). Os corpos-de-prova foram caracterizados antes e após os ensaios de fadiga térmica. Antes dos ensaios, caracterizaram-se os carbonetos eutéticos (tipos, morfologia, fração volumétrica, tamanho, forma e distribuição dos carbonetos) e a microdureza da matriz. Após os ensaios, caracterizaram-se as trincas de fadiga térmica macroscópicas e microscópicas (número e profundidade) e a microdureza da matriz. A nucleação de trincas de fadiga térmica ocorre predominantemente na superfície do corpo-de-prova, induzidas por concentradores de tensão mecânicos e metalúrgicos. As trincas nucleiam na matriz (a rugosidade age como concentrador de tensão mecânico) e em carbonetos (interface carboneto/matriz ou no próprio carboneto). A taxa de nucleação sofre influência da fração volumétrica de carboneto eutético (seu aumento promove aumento da taxa de nucleação) e da dureza da matriz (seu aumento promove diminuição da taxa de nucleação). A propagação de trincas ocorre predominantemente pela interface carboneto/matriz ou através do carboneto. A taxa de propagação sofre influência da distribuição de carboneto eutético. Quanto maior a relação “continuidade de carbonetos/distância livre média entre carbonetos”, maior a taxa de propagação de trincas. A taxa de propagação de trinca diminui com o aumento do tempo de ensaio, independentemente da fração volumétrica de carboneto eutético e da dureza da matriz. O regime de propagação durante os primeiros 100 ciclos é caracterizado pela propagação instável da trinca controlada pela tenacidade à fratura do material; de 100 a 500 ciclos, a propagação é controlada pela magnitude da tensão. O tamanho do corpo-de-prova também influenciou os resultados dos ensaios: o aumento do tamanho promove aumento das taxas de nucleação e propagação. Este resultado é atribuído ao aumento do gradiente térmico ao longo do corpo-de-prova com o aumento do seu diâmetro. / The effects of the volume fraction of eutectic carbides and of the matrix hardness on the thermal fatigue resistance of multicomponent white cast iron were investigated. Alloys Fe-4Cr-V-2Mo-2W-2C, V ranging from 5 to 8 wt% and Fe-4Cr-8V-Mo-2W-2C, Mo ranging from 2 to 5 wt % were used. Disc shaped samples were quench and tempered for obtaining two matrix microhardness levels: 450 HV and 650 HV. Thermal fatigue tests were carried out for 100 and 500 cycles. Each cycle involved high frequency induction heating of the surface to 600°C and subsequent cooling in water during 45 seconds (equalization of the bulk and surface temperature). The test specimens were characterized before and after the thermal fatigue tests. Before the tests, eutectic carbide (type, morphology, volume fraction, syze, shape and distribution of carbides) and matrix microhardness were characterized. After the tests, the macroscopic and microscopic thermal fatigue cracks (number and depth) and matrix microhardness were characterized. The nucleation of the thermal fatigue cracks takes place mostly at the specimen surface, induced by mechanical and metallurgical stress risers. The crack nucleates at the matrix (roughness as mechanical stress risers as well as at carbides (at the carbide/matrix interface or at the carbide itself). The nucleation rate is influenced by the volume fraction of eutectic carbide (the higher the volume fraction, the higher the nucleation rate) and by the matrix microhardness (the higher the microhardness, the lower the nucleation rate). The crack propagation mostly takes place at the carbide/matrix interface or through the carbide. The propagation rate is affected by the carbide distribution. The higher the “carbide continuity/carbide free path” ratio, the higher the propagation rate. The propagation rate decreases with increasing test time, regardless the eutectic carbide volume fraction and the matrix microhardness. The propagation behaviour during the first 100 cycles is characterized by instable crack propagation controlled by the fracture toughness of the material; from 100 to 500 cycles, the propagation is controlled by the stress magnitude. The syze of the test specimen also influenced the tests results: the larger the specimen syze, the higher the nucleation and propagation rates. This is attributed to the effect of increasing thermal gradient across the specimen with increasing specimen diameter.

O presente trabalho teve como objetivo principal estudar a evolução da microestrutura e da textura cristalográfica de três alumínios de pureza comercial com diferentes níveis de pureza, AA1100 (99,00%), AA1050 (99,50%) e AA1070 (99,70%), e do alumínio super puro, AA1199 (99,995%). A obtenção das placas de alumínio de pureza comercial foi feita pelo processo Direct Chill (DC) seguido de homogeneização e laminação a quente até a espessura final de 10 mm e o alumínio super puro foi produzido em laboratório e encontrava-se na forma de blocos fundidos. Foram utilizadas várias técnicas de análise microestrutural: microscopia óptica convencional, microscopia óptica com luz polarizada, microscopia eletrônica de varredura, microscopia eletrônica de transmissão, medidas de condutividade elétrica e dureza Vickers. Já a evolução da textura do material foi analisada pela técnica de difração de raios X, nos alumínios de pureza comercial e pela técnica de difração de elétrons retroespalhados (EBSD), no alumínio super puro. No estado como recebido e recozido a 400 ºC por 1 e 24 horas foi possível observar que nos alumínios de pureza comercial produzidos pelo processo DC não há variações na composição química ao longo da espessura. Os precipitados nos alumínios AA1100 e AA1070, no estado como recebido estavam distribuídos na direção de laminação, porém com vários locais de aglomeração dos mesmos, e no alumínio AA1050 os mesmos estavam finamente dispersos na direção de laminação. Após recozimento a 400 ºC por 1 e 24 horas as mesmas características foram mantidas, porém ocorreu um aumento na fração volumétrica do estado como recebido para os recozidos a 400 ºC, ou seja, ocorreu precipitação durante o recozimento. Quanto a análise da textura dos alumínios de pureza comercial, na superfície encontrou-se a textura de cisalhamento, devido ao esforço mecânico entre o cilindro de laminação e as amostras, tanto no estado como recebido como após os recozimentos. A ¼ da espessura e no centro do material após recozimentos ainda observou-se a textura de cisalhamento, mas também as texturas tipo cubo, cubo rodado e latão. No centro do material permaneceram a textura tipo cubo e latão. Já o alumínio AA1199 possui uma estrutura grosseira (grãos oligocristalinos), sem a ocorrência de precipitados e não foi possível obter resultados nas análises de textura, devido ao tamanho de grão grande. Após a laminação a frio com aproximadamente 70% de redução em espessura, seguido de tratamentos isócronos no tempo de 1 hora e nas temperaturas de 50, 100, 150, 200, 250, 300, 350, 400 e 500 ºC, o alumínio AA1199 apresentou uma menor temperatura para completa recristalização, em torno de 280 ºC, seguido do alumínio AA1050 (375 ºC) e dos alumínios AA1100 e AA1070 (ambos, 400 ºC). Como as composições químicas das soluções sólidas matrizes são praticamente idênticas, a variação na temperatura de recristalização deve ser atribuída aos precipitados presentes no alumínio que quando finamente dispersos na matriz impedem o movimento dos subcontornos, retardando o início da recristalização. O alumínio AA1199 possui uma temperatura para completa recristalização, menor, em relação aos outros alumínios, uma vez que não possui precipitados em sua microestrutura. Quanto a textura, na superfície e a ¼ da espessura, os alumínios de pureza comercial laminados a frio sem tratamento térmico, permaneceram com a textura de cisalhamento encontrada na superfície no estado como recebido com o surgimento da textura S. Após o tratamento térmico a 250 ºC por 1 hora apareceu a textura Goss e a 350 ºC por 1 hora a textura tipo cubo, cubo rodado e cubo ND, lembrando que em ambos os casos a textura de cisalhamento e S permaneceram. Com o auxílio da utilização da técnica de difração de elétrons retroespalhados (EBSD) encontrou-se no alumínio AA1199, a textura tipo cubo, tanto laminado a frio sem tratamento térmico quanto após os tratamentos. Após a laminação a frio com aproximadamente 80% de redução em espessura, seguida de tratamentos isotérmicos, nas temperaturas de 150, 250 e 350 ºC e nos tempos de ½, 1, 2, 3 e 4 horas, os alumínios de pureza comercial apresentaram uma temperatura para completa recristalização em torno de 350 ºC e por meio da técnica de microscopia eletrônica de transmissão observou-se arranjos celulares de discordâncias, mas não foram detectadas diferenças significativas entre esses arranjos apresentados pelos três alumínios. Diferenças certamente existem, mas para serem caracterizadas é necessário um estudo detalhado com microscopia eletrônica de transmissão. A presença de partículas de compostos intermetálicos foi observada em todas as amostras analisadas, essas partículas são resultado da baixa solubilidade do ferro e do silício no alumínio e se formaram durante as diversas etapas do processamento dos materiais. / The present work had as main objective to study the evolution of the microstructure and crystallographic texture of three commercial purity aluminums with different levels of purity, AA1100 (99.00%), AA1050 (99.50%) and AA1070 (99.70%), and of the super pure aluminum, AA1199 (99.995%). The plates of commercial purity aluminum were obtained through direct chill (DC) process followed by homogenization and hot rolling until the final thickness of 10 mm. The super pure aluminum was produced in laboratory and was in the ascast condition. Several techniques of microstructural analysis were used: conventional optical microscopy, optical microscopy using polarized light, scanning electron microscopy, transmission electron microscopy, measurements of electrical conductivity and Vickers hardness. The evolution of the texture of the commercial purity aluminum was analyzed by Xray diffraction and the super pure aluminum was analyzed using electron backscatter diffraction. Variations in the chemical composition along thickness in the commercial purity aluminum produced by the DC process in the as-received condition and after annealing at 400 ºC for 1 and 24 hours were possible to observe. The precipitates in the aluminum AA1100 and AA1070, in the as-received condition were distributed in the rolling direction, including several sites of clusters of the same precipitates. In the aluminum AA1050 the precipitates were finely dispersed in the rolling direction. After annealing at 400 ºC for 1 and 24 hours the precipitate characteristics were maintained, and there was an increase in the volumetric fraction of the as-received condition in comparison to samples annealed at 400 ºC, showing that precipitation occurred during annealing. The shear texture was observed in the commercial purity aluminums, either in the as-received condition or after annealing, and this effect was due to the mechanical effort between the rolling mill and the samples. At ¼ of the thickness and in the center of the material after annealing the shear texture was still observed, and also the type cube, rotated cube and brass textures. The center of the material continues showing the type cube and brass textures. On the other hand, the aluminum AA1199 has a coarse structure (oligocrystalline grains), without the occurrence of precipitates, and texture results were not obtained due to the fact the grains are much coarser, due to the size of big grain. After cold rolling with approximately 70% of reduction in thickness, followed by isochronous treatments for 1 hour at temperatures of 50, 100, 150, 200, 250, 300, 350, 400 and 500 ºC, the aluminum AA1199 presented a smaller temperature for complete recrystallization, around 280 ºC, followed by the aluminum AA1050 (375 ºC) and of the aluminum AA1100 and AA1070 (both, 400 ºC). As the chemical compositions of the solid solutions matrixes are almost identical, the variation in the recrystallization temperature should be attributed to the presence of precipitates in the aluminum, and when finely dispersed in the matrix they impede the movement of the sub-boundary, delaying the beginning of the recrystallization. In comparison to the other aluminums, the aluminum AA1199 has a smaller temperature for complete recrystallization, since there are not precipitates in its microstructure. Observing the texture, in the surface and the ¼ of the thickness, the commercial purity aluminums, after cold rolling and without heat treatment, showed the shear texture found in the surface in the as-received condition with the appearance of the S texture. After heat treatment at 250 ºC for 1 hour the Goss texture appeared and at 350 ºC for 1 hour the type cube, rotated cube and cube ND textures; notice that in both cases the shear and S texture were still present. Energy dispersive X-ray spectroscopy (microanalysis) showed in the aluminum AA1199, the texture type cube, so cold rolled without heat treatment as after the treatments. After cold rolling with approximately 80% of reduction in thickness, followed by isothermal treatments, at temperatures of 150, 250 and 350 ºC and times of ½, 1, 2, 3 and 4 hours, the commercial purity aluminum presented a temperature for complete recrystallization around 350 ºC. Transmission electron microscopy technique showed dislocations cellular arrangements, but significant differences were not detected among those arrangements presented by the three aluminums. Differences certainly exist, but for their characterization detailed study with transmission electron microscopy is necessary. The presence of particles of intermetallic compounds were observed in all analyzed samples, those particles resulted of the low solubility of iron and silicon in the aluminum and were formed during the several stages of the processing of the materials.

O aço 10B06 é utilizado na produção de fio máquina para trefilação. Por ser produzido em aciaria elétrica, o teor de nitrogênio desse aço tende a ser elevado, o que pode causar o fenômeno de envelhecimento, resultando em aumento de resistência e queda de ductilidade. Reduzir o teor de nitrogênio livre é uma forma de minimizar esse efeito, o que é realizado pela adição de boro. Dessa forma, o presente trabalho teve como objetivo estudar o efeito do boro no aço 10B06. Foram utilizadas amostras de fio máquina, diâmetro 5,50mm, e arame trefilado com teores de boro variando de 55 a 112ppm. Os dois principais fatores que influenciam a microestrutura e propriedades mecânicas do material são a composição química e o processo produtivo. Por esse motivo, inicialmente foi avaliada a etapa de resfriamento durante a laminação do fio máquina, por meio de uma análise termográfica. Foi encontrada a temperatura de 767ºC no início do resfriamento do material, valor próximo de Ac1, limite inferior do intervalo intercrítico. Esse fato favorece o enriquecimento da austenita em carbono, facilitando a formação de martensita e/ou austenita retida. Os resultados da análise metalográfica com Nital revelaram microestrutura composta de aproximadamente 88% de ferrita e 12% de perlita. Já o reagente Le Pera identificou ilhas de martensita e/ou austenita retida (MA), o que foi comprovado por análises de microscopia eletrônica de varredura. Os resultados de microdureza mostraram uma tendência de queda com o aumento do teor de boro. As medições de nanodureza confirmaram a presença do microconstituinte MA. A ferrita apresentou dureza na faixa de 280 - 320 HV, enquanto o MA ficou no intervalo de 349 - 424 HV. Comparando-se o limite de resistência obtido no fio máquina com menor teor de boro (55ppm - 434MPa) com o de maior (112ppm - 394MPa), observasse uma redução de 40MPa para um acréscimo de 57ppm de boro, confirmando o efeito esperado para esse elemento pela formação de BN. Utilizando os dados da curva tensão versus deformação foram obtidos os coeficientes de encruamento dos fios máquina, que apresentaram uma tendência de crescimento com o teor de boro, sendo que o material com 101ppm de boro mostrou o melhor valor para aplicação em processos de conformação. As análises fractográficas revelaram que os mecanismos de fratura presentes em todos os materiais foram a formação e coalescimento de microvidades. Considerando-se todas as análises realizadas nesse trabalho foi possível evidenciar o efeito do boro na redução do efeito do envelhecimento provocado pelo nitrogênio. Entre os testes realizados, os que apresentaram as melhores relações B/N foram os de teores 109 e 112ppm de boro. Essas relações estão próximas do valor máximo indicado de 0,8. / The 10B06 steel grade is applied to the production of wire rod for drawing. The steel is produced in an Electric Arc Furnace and this process can lead to higher nitrogen levels that cause strain aging resulting in an increase of the strength and ductility loss. The reduction of free nitrogen content through boron addition is one way to minimize this effect that can be done through boron addition. In this way, the present work aims to study boron effect on 10B06 steel. It was used wire rod, diameter 5,50mm and a cold-drawn wire rod with chemical composition from 55 to 112ppm of boron. The main factors that influence the microstructure and mechanical properties of the material are the chemical composition and the productive process. For this reason, initially, the cooling treatment was analyzed using thermography and it was found at the beginning of the cooling process a temperature of 767oC in the material, close to Ac1, lower limit of the intercritical interval. This temperature can facilitate the enrichment of carbon in the austenite, making easier martensite and/or retained austenite formation. The metallographic analysis with Nital etchant showed a microstructure composed of about 88% ferrite and 12% perlite, and Le Pera etchant identified martensite and/or retained austenite (MA), what was confirmed by scanning electron microscope analysis. Furthermore, nanohardness measurements confirmed the presence of the microconstituent MA, as ferrite presented hardness from 280 to 320 HV and MA from 349 to 424 HV. Microhardess results decreased with increasing boron content. A decrease of 40MPa in the tensile strength was observed between the wire rod with boron content of 55ppm (434MPa) and the one with 112ppm (394MPa), in agreement with the expected effect of the BN formation. It was possible to obtain hardening coefficient of the wire rod using the data of the stress versus strain curve, which increased with boron content, and the material with 101ppm of boron showed the best value for forming application. Fractographic analysis revealed that the fracture mechanisms in all the materials were the development and coalescing of microcavity. Considering all the analysis of this study it was possible to confirm boron\'s effect to reduce strain aging caused by nitrogen. Materials with 109 and 112ppm of boron presented the best B/N ratio, with values close to the maximum indicated, 0,8.

Embora a literatura disponível sobre conformação mecânica disponha de grande número de trabalhos relacionados à fadiga térmica em cilindros de laminação a quente, poucos autores abordam como esse tipo de dano é influenciado por algumas características do processo - em especial a oxidação na superfície de trabalho dos cilindros, tanto enquanto formação de uma camada de óxido uniforme sobre essa superfície quanto na forma de corrosão. Essa lacuna ocorre a despeito do volume de referências que relacionam a oxidação à fadiga térmica em escala laboratorial. O presente trabalho busca relacionar ambas as linhas de investigação, utilizando métodos analíticos e numéricos comparados com experimentos em diferentes escalas, para compreender sob quais condições são formados os óxidos presentes nos cilindros, qual seu comportamento mecânico no decorrer dos ciclos de laminação e como os óxidos influenciam na ocorrência de trincas de laminação a quente. A análise da superfície de trabalho de um cilindro em escala piloto revela a presença de trincas térmicas na superfície de trabalho. Os padrões de defeitos obtidos se mostram de acordo com o exposto em diversas referências publicadas sobre ensaios de fadiga térmica e termomecânica, em relação à presença e distribuição de óxidos no interior das trincas e na superfície externa das amostras. Em paralelo, foram desenvolvidos modelos numéricos do processo de laminação a quente, inicialmente em escala macroscópica, para obter detalhes sobre distribuições de temperatura, tensões e deformações nos cilindros, de difícil obtenção experimental. Estes modelos foram simulados e seus resultados comparados com valores medidos no laminador piloto, validando seu uso para avaliação do comportamento dos cilindros durante uma série de campanhas de laminação. Os resultados obtidos mostram que o mecanismo de fadiga de baixo ciclo normalmente associado ao dano térmico não é suficiente para explicar a nucleação das trincas de fadiga térmica observadas no experimento. O erro na predição pela teoria de fadiga de baixo ciclo tradicional, em conjunto com a caracterização dos cilindros em escala piloto e com observações de cilindros em escala industrial presentes na literatura, permite levantar a hipótese de que os fenômenos de oxidação e corrosão em curso durante o processo de laminação contribuem para uma redução significativa da resistência dos cilindros à fadiga termomecânica. Cálculos analíticos e simulações adicionais em escala microscópica foram realizados, incluindo os efeitos da oxidação e da corrosão, confirmando a hipótese e oferecendo uma linha de investigação mais assertiva para a nucleação e evolução do dano nos cilindros. / Although the literature on metal forming presents a considerable amount of references related to thermal fatigue in hot rolling, only few authors discuss how this type of damage is influenced by some characteristics of the rolling process - specially the oxidation that develops at the surface of the rolls. This gap occurs despite the amount of references relating oxidation to thermal fatigue in laboratory scale. The current work aims to relate both investigation paths, employing analytical and numerical methods in comparison with experiments in different scales, to provide understanding about the conditions at which the oxides observed upon the rolls are formed, their mechanical behaviour during the rolling process and how they influence thermal fatigue cracking in the rolls. The analysis of the roll surface in a pilot scale mill reveals the presence of thermal fatigue cracks on the work surface. The patterns of defects observed are in accordance with several references on thermal and thermomechanical fatigue, regarding the presence and distribution of oxides inside the cracks and at the external surface of the samples. In parallel, numerical models of hot rolling were developed, initially at macroscopic scale, to obtain details on temperature, stress and strain distribution in the rolls, which are difficult to obtain experimentally. These models were simulated and their results were compared with values measured in the pilot scale mill, validating their use to evaluate the behaviour of the rolls throughout a series of rolling campaigns. The obtained results show that the low-cycle fatigue mechanism commonly associated to thermal damage is not enough to explain the nucleation of thermal cracks seen in the experiments. The error in the prediction using the traditional low-cycle fatigue theory, in conjunction with the characterization of the pilot-scale rolls and with observations of industrial scale rolls, yields the hypothesis that oxidation and corrosion phenomena occurring throughout the rolling process contribute to a significative reduction in the resistance of the rolls to thermomechanical fatigue. Analytical calculations and additional simulations in microscopic scale were conducted, including effects of oxidation and corrosion, confirming the hypothesis and providing a more assertive line of research for damage nucleation and evolution in the rolls.

Orientador: Sérgio Tonini Button / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-20T23:55:31Z (GMT). No. of bitstreams: 1 Judice_GustavoHenrique_M.pdf: 13713684 bytes, checksum: ca6f8c3e7b2f080bdd41cef45bd2a552 (MD5) Previous issue date: 2012 / Resumo: O processo de conformação denominado Laminação Transversal com Cunha, industrialmente conhecido como Cross Wedge Rolling (CWR), tem se destacado na indústria pela sua flexibilidade, produtividade e economia de material, porém pode apresentar defeitos internos que, em alguns casos, inviabiliza a utilização dos produtos. É um defeito de difícil avaliação quanto a sua formação e propagação e ainda não se desenvolveu um mapa de processamento que represente de forma precisa e consistente as condições em que o defeito pode aparecer. Este estudo apresenta resultados dos campos de deformação plástica, tensão, temperatura e do aspecto do desenvolvimento de defeitos internos, acrescentados ao tarugo, através de simulações do processo de laminação transversal com cunha. O estudo se desenvolveu na criação de modelos de elementos finitos pelo software comercial Abaqus® versão 6.9-2, que seguiram condições geométricas de um processo real. Inicialmente, modelos bidimensionais foram simulados apresentando diferentes densidades de malha. Posteriormente, modelos com um vazio central foram simulados e confirmou-se a boa correlação de seus resultados com os encontrados na literatura. Na sequência, um modelo com propriedades de porosidade indicou que a região central do tarugo apresentou a maior densidade de vazios. Também foram simulados modelos com vazios e inclusões de diferentes tamanhos, disposições e formas, com destaque para a região em volta das inclusões, por apresentar a formação de vazios. Por último, foram simulados dois modelos tridimensionais, um apresentando um vazio interno no centro do tarugo, que apresentou os campos de tensão, de deformação plástica e de temperatura parecidos com o do modelo bidimensional correspondente, porém com valores maiores. O outro modelo tridimensional apresentou propriedades de porosidade e indicou que a maior probabilidade de formação de defeitos se concentrou na extensão longitudinal central do tarugo. Os maiores níveis de temperatura, deformação plástica e tensão foram presentes na região central do tarugo, tanto nos modelos bidimensionais quanto nos tridimensionais, indicando que essa região apresentou os maiores níveis de solicitação, o que pode favorecer o surgimento dos defeitos internos / Abstract: Cross Wedge Rolling is a metal forming process which has become very important in the industry for its flexibility, productivity and material savings, although presenting the critical problem of internal defects generation which restricts the use of rolled products, since it is very difficult to asset the defects formation and propagation and it was still not found a processing map to represent CWR with precise and consistent conditions with which the defect should or should not appear. In this study some numerical results are presented in terms of plastic strain, stress, temperature and the development of internal defects within the billet. These results were obtained with different finite element models developed with the commercial software Abaqus Version 6.9-2 by simulating materials with different internal characteristics, representing geometric conditions identified in the actual process. Initially two-dimensional models were simulated with different mesh densities; then, models with a central void were simulated and a good relationship was found with the results found in the literature. Then a model was simulated to characterize the porosity of the material, and was verified that the porosity is concentrated at the center of billet. Other models were also simulated with voids and inclusions of different sizes, distributions and shapes which were added manually. These models showed similar results, and the models with inclusions developed voids within the metallic matrix. Finally, two models (three-dimensional) were simulated, one with an internal void in the center of the billet, which showed the development of stress fields, plastic deformation and temperature similar to that corresponding two-dimensional model, but with higher values. The other three-dimensional model, which presented properties of porosity, showed that the highest probability of defect formation focused in the central and longitudinal extent region of the billet. All models showed that the central region of the billet presents the highest levels of temperature, plastic strain and stress probably favoring the initiation of internal defects / Mestrado / Materiais e Processos de Fabricação / Mestre em Engenharia Mecânica

[pt] Ondas superficiais são uma fonte comum de ruído nos dados sísmicos. O rolamento superficial pode ser definido como um conjunto de eventos que contaminam os dados sísmicos formando um cone de alto mergulho (baixa velocidade) e alta amplitude, podendo dominar as componentes do sinal que estão perto da fonte. Ondas de superfície são um problema tão significativo na aquisição de dados terrestres que no desenvolvimento dos parâmetros de aqusição é levado em consideração a necessidade de suprimí- las. Este trabalho propõe a remoção do ruído de rolamento superficial a partir da transformada wavelet em uma dimensão. Este método obteve melhores resultados que a transformada f-k. O algoritmo foi testado em dados sintético e real. / [en] Surface waves are a common source of unwanted noise in the seismic record. Ground roll can be defined as a group of events that contaminates seismic data and makes a high dip (low velocity) and high amplitude cone, this can dominate near-source on seismic records. Surface waves are such a problem in land seismic record acquisition that the design of acquisition parameters is dominated by the need to suppress them. In this work we propose to remove ground roll by employing Wavelet Transforms in 1-D. This method gives better results than the f-k transform. The procedure was tested in synthetic and real data.

Background: Nowadays, one of the medical problem is rolling cell adhesion. Rolling cell adhesion is a complex process that requires the analysis of the challenging environment such as body fluid and is the process responsible for recruiting the cell to specific organs. In order to explore the rolling cell adhesion, mathematical model is proposed. Different image processing methods are created, such as optical flow - Lucas Kanade algorithm, and other type of methods related to mechanical fluid, namely PIV (Particle Image Velocimetry). Aim: The aim of this master thesis is the identification of challenges while using PIV in live cell images and propose the algorithm, which may analyze the rolling cell adhesion problem. Methods: In order to understand properly the rolling cell adhesion problem from biological site, literature review combined with the expert consultation is performed. According to gather information, mathematical model is proposed. Particle Image Velocimetry is explained according to literature review, where at the beginning the expert recommends some books as a primary research. As a result of this research, PIV challenges are identified and generally PIV idea is explained. Then two experiments are performed. The first experiment evaluates detection algorithms and the second one, analyses track algorithm vs. PIV. In order to evaluate the mentioned algorithms, some evaluation method are selected and some criteria are defined. Unfortunately the found methods are not perfect, therefore a new method related to performance evaluation using time series is proposed. Thesis result: The result of this thesis is a proposition of the algorithm, which can be used in the rolling cell adhesion. The algorithm is formed according to the detailed exploration of the rolling cell adhesion and analysis of the selected algorithms related to the image analysis during the theoretical research and experiments.

This work has been performed at Outokumpu Stainless AB, QPE, in Degerfors where focus is on production of hot rolled tailor-made quarto plate in special stainless steel grades such as highly alloyed austenitics and duplex steel grades. After a period with low values in strength and a lot of deviations in tensile tests for the highly alloyed austenitic stainless steel, 254 SMO, there was a need to investigate what influence some parameters during process has on strength values. Process parameters during hot rolling and levelling have been examined to see if there is a connection between these and the strength. A statistical data base has been created to investigate the connection with strength. There has been experimental work involving two test plates in different dimensions to see how a double/extended heat treatment influence the sigma phase values and grain size, and if these can be connected to the strength values. The test plates have also been used to investigate how the position of the test coupon is affecting the tensile test result. Moreover, there have been tests performed in the cold plate leveller. The strength is a complex problem since there are many parameters that contribute to a variation in strength value. It is possible to see an influence from the investigated parameters, but their individual impact on strength varies. There are several indications in the results that the flatness seems to be a factor affecting the variation in strength. However, there is a need for more thorough investigations where the flatness can be measured to completely verify this / Det här arbetet har utförts på Outokumpu Stainless AB, QPE, i Degerfors där man inriktar sig på produktion av varmvalsad, skräddarsydd kvartoplåt i olika specialstål som höglegerade austeniter och duplexa stålsorter. Efter en period med låga värden i hållfasthet och många avvikelser i provning för ett höglegerat austenitiskt stål, 254 SMO, så fanns ett behov av att undersöka vissa parametrars inverkan på hållfastheten. Processparametrar vid valsning och riktning har undersökts för att se om det går att se ett samband mellan dessa och hållfastheten. Statistisk data har tagits fram för att undersöka sambandet med hållfastheten. Även experimentella försök har gjorts, främst på två testplåtar i olika dimensioner, för att se hur en dubbel/förlängd värmebehandling påverkar sigmafas och kornstorlek men även om dessa värden, genom försöken, kan kopplas till hållfasthetsvärden. Plåtarna har också använts till att se hur testresultatet påverkas av var i plåten som testkupongen tas ut. Dessutom har det gjorts försök i riktverket med olika parametrar. Hållfastheten är ett komplext problem att utreda eftersom det finns så många parametrar som ger en variation i resultaten. Det är möjligt att se en påverkan från flera av de undersökta parametrarna men deras individuella påverkan på hållfastheten varierar. Det finns många indikationer i resultaten som pekar på att planheten verkar vara en faktor som påverkar den variation som finns i resultat. Dock behövs det grundligare undersökningar där planheten kan mätas för att helt verifiera detta.

Abstract In today's world, enormous amounts of data are being collected everyday. Thus, the problems of storing, handling, and utilizing the data are faced constantly. As the human mind itself can no longer interpret the vast datasets, methods for extracting useful and novel information from the data are needed and developed. These methods are collectively called knowledge discovery methods. In this thesis, a novel combination of feature selection and data modeling methods is presented in order to help with this task. This combination includes the methods of basic statistical analysis, linear correlation, self-organizing map, parallel coordinates, and k-means clustering. The presented method can be used, first, to select the most relevant features from even hundreds of them and, then, to model the complex inter-correlations within the selected ones. The capability to handle hundreds of features opens up the possibility to study more extensive processes instead of just looking at smaller parts of them. The results of k-nearest-neighbors study show that the presented feature selection procedure is valid and appropriate. A second advantage of the presented method is the possibility to use thousands of samples. Whereas the current rules of selecting appropriate limits for utilizing the methods are theoretically proved only for small sample sizes, especially in the case of linear correlation, this thesis gives the guidelines for feature selection with thousands of samples. A third positive aspect is the nature of the results: given that the outcome of the method is a set of conditional probabilities, the derived model is highly unrestrictive and rather easy to interpret. In order to test the presented method in practice, it was applied to study two different cases of steel manufacturing with hot strip rolling. In the first case, the conditional probabilities for different types of retentions were derived and, in the second case, the rolling conditions for the occurrence of wedge were revealed. The results of both of these studies show that steel manufacturing processes are indeed very complex and highly dependent on the various stages of the manufacturing. This was further confirmed by the fact that with studies of k-nearest-neighbors and C4.5, it was impossible to derive useful models concerning the datasets as a whole. It is believed that the reason for this lies in the nature of these two methods, meaning that they are unable to grasp such manifold inter-correlations in the data. On the contrary, the presented method of conditional probabilities allowed new knowledge to be gained of the studied processes, which will help to better understand these processes and to enhance them.

M.Ing. / The aim of Spoornet is to provide a minimise predictable service. In order to provide a predictable service, it is necessary to move trains safely and effectively from the place of departure to their destination. The keywords here are safely and effectively. Although support functions such as infrastructure and train control procedures are vital in moving the train, the train or rolling stock as it is generally known, warrants some attention. Defects on the rolling stock are very costly to Spoornet. This is mainly due to the fact that a defect on the rolling stock that goes undetected can cause damage to the rolling stock and the infrastructure. This damage can eventually lead to derailments. Considering that a derailment can cost Spoornet millions of rand and cause delays to the services, it is only logical to spend time preventing derailments. It is for this reason that a workgroup was formed to investigate and solve the problem of defects causing derailments and delays by developing an early warning system. The need for an integrated train condition monitoring system became apparent when considering an early warning system. The objectives of the integrated train condition monitoring system are to provide train condition information to different users, and alarms on detection of emergency or dangerous conditions. Various train defects that may cause damage or derailments were identified. One of them being a flat wheel on a rail vehicle. A flat wheel is characterised by the flattening of the wheel on one or more positions on its circumference, so that the wheel does not have an even and completely round profile. Flat wheels are mostly caused by the wheels of a vehicle becoming locked during braking, and sliding along the rail track. The friction created by this action grinds a flat spot on the wheel. The flat wheel leads to a decline in the riding quality of the rolling stock and a rise in the levels of vibration and noise is evident. But more importantly, the flat spot causes the wheel to roll unevenly, creating impacts on the rail on some points. It is these impacts that can cause damage to the rail and the rolling stock. Depending on the length of the flat spot, the vehicle type and speed, the stresses may be sufficient to cause final failure of the rail or initiate fatigue cracks in the rail. Severe flat wheels are a safety hazard and can in some cases, cause derailments and consequent delays to trains. Smaller flat spots contribute to track deterioration and so increase maintenance costs by damaging the rails, sleepers and ballast. Flat wheels can thus be very costly to Spoornet and its public image. In addition to safety and economic considerations, wheel flats reduce the comfort levels in the passenger coaches and the noises they make is annoying. In an attempt to restrict the damage caused by flat wheels, most railway administrations place a limit on the length of the flats that may stay in service. But to effectively find a flat wheel on rolling stock is currently a very expensive exercise. Flat wheels can be detected by an audible knocking sound when standing next to the rail. This sound is impossible for the driver to hear and therefore goes undetected. Normally flat wheels are detected by random inspection of the rolling stock or when they are brought in for a routine service. The service cycle on rolling stock can be up to 24 months in Spoornet. Considering that a flat wheel has an impact roughly every 3m, a serious flat generates roughly 160 000 impacts on a single trip on the coal heavy haul export line. It is therefore clear that a flat wheel can cause a considerable amount of damage between service cycles. The severity of the problem is however not accurately defined in Spoornet, because up to a few months ago there was no detection system in use to determine the distribution of flat wheels. The research department of the Deutsche Bundesbahn however considers rail fractures due to the flat wheels to be a serious problem with a significant annual replacement cost. There are thus sound safety and economic reasons for wishing to understand the mechanisms of flat wheels and to develop an early warning system using an automatic detector.