A presente tese de doutorado avalia a Emissão Magnética Espontânea (SME), fenômeno descrito recentemente no grupo de pesquisa do professor Hélio Goldenstein e que já demonstrou ser uma ferramenta promissora para monitorar a transformação martensítica nos aços. O SME é utilizado para monitorar a velocidade de propagação de plaquetas de martensita; identificar a temperatura em que ocorre o inicio da primeira transformação (Ms); observar martensita induzida por deformação ou tensões durante a deformação de aços contendo austenita metaestável e também estudar a transformação isotérmica da austenita em martensita. A duração dos picos individuais magnéticos emitidos no início (Ms) e no final (Mf) da transformação martensítica durante o resfriamento contínuo foram medidos e comparados com os tamanhos médios das placas de martensita, obtidos através da metalografia quantitativa. O SME, um fenômeno mais sensível do que outras medidas globais como a resistividade, dilatometria, Ruído Magnético de Barkhausen (MBN), etc., não exige um volume crítico de transformação para ser detectado, e é capaz de detectar sinais de placas individuais crescentes. / This PhD thesis evaluates the Spontaneous Magnetic Emission (SME), a phenomenon recently described in the research group of Professor Hélio Goldenstein that has proved to be a promising tool to monitor the Martensitic Transformation in steels. The SME is used to monitor the speed of propagation of martensite plates; to identify the temperature at which the first transformation occurs (Ms); to observe martensite straininduced or stresses during deformation of steels containing metastable austenite and also study the isothermal transformation of austenite in martensite. The duration of individual magnetic peaks issued at the start (Ms) and in the end (Mf) of the martensitic transformation during the continous cooling were measured and compared with the average size of the martensite plates obtained by quantitative metalography. The SME, a phenomenon more sensitive than other global measures such as resistivity, dilatometry, Magnetic Barkhausen Noise (MBN), etc., does not require a critical volume of transformation to be detected, and is able to detect signs of growing individual plates.

No presente trabalho foram estudados os fenômenos de encruamento e, principalmente, a formação e reversão da martensita alfa-linha (a\', cúbica de corpo centrado, CCC, ferromagnética) induzida por deformação em um aço inoxidável dúplex UNS S31803 e um super dúplex UNS S32520. Inicialmente, as microestruturas dos dois materiais na condição solubilizada foram caracterizadas com auxílio de várias técnicas complementares de análise microestrutural. Foram determinadas fração volumétrica, estrutura cristalina, composição química, tamanho e morfologia das duas fases (ferrita e austenita). Posteriormente, os dois aços foram deformados por dois métodos: a laminação a frio, dividida em vários estágios, com menores graus de deformação e a limagem, sendo que o cavaco limado resultante apresenta altos graus de deformação. Algumas amostras deformadas foram recozidas. Os fenômenos de encruamento, formação e reversão de martensita induzida por deformação na austenita, recuperação, recristalização da austenita e da ferrita no cavaco limado foram estudados predominantemente por difratometria de raios X e usando o método de Rietveld. A difratometria de raios X também foi utilizada para determinação das microdeformações residuais e tamanhos de cristalito (subgrão), calculadas a partir do alargamento dos picos de difração causado pelas deformações. Desta forma, puderam-se comparar os níveis de deformação da laminação e limagem. Qualitativamente, a formação e reversão da martensita induzida por deformação também foi estudada por meio de medidas magnéticas utilizando-se dados de saturação magnética das curvas de histerese obtidas com o auxílio de um magnetômetro de amostra vibrante. Observou-se que para o aço inoxidável dúplex, tanto a laminação quanto a limagem causaram a formação de martensita induzida por deformação e para o aço inoxidável super dúplex, apenas a limagem promoveu essa transformação. Em comparação com o aço dúplex, o aço super dúplex apresentou maior resistência à formação de martensita induzida por deformação, pois apresenta uma austenita mais rica em nitrogênio e uma maior propensão à formação de fase sigma durante o recozimento, pois apresenta uma ferrita mais rica em cromo e nitrogênio. / In the present work the phenomena of strain hardening, formation and reversion of the strain induced alpha-prime martensite (a\', body centered cubic, BCC, Ferromagnetic) in an UNS S31803 duplex and UNS S32520 super duplex stainless steels have been studied. Firstly, the microstructures of both materials in the solution annealed condition were characterized with the aid of several microstructural analysis complementary techniques. The volume fraction, crystalline structure, chemical composition, size and morphology of the two phases (ferrite and austenite) have been determined. Further, both steels were deformed by two methods: cold rolling, divided into several stages, with lower strain levels than filing, which the chips resulting had higher strain levels. The phenomena of strain hardening, formation and reversion of strain induced martensite in the austenite phase, recovery and recrystallization of austenite and ferrite phases have been studied, mainly using X-ray diffraction and the Rietveld method. X-ray diffraction was also used to determine the residual microstrain and crystallite size (sub grain), calculated from the diffraction peak broadening caused by straining. Thus, the levels of cold rolling and filing strains could be compared. Qualitatively, the formation and reversion of strain induced martensite was also studied by magnetic measurements using data from magnetic saturation of hysteresis curves obtained with the aid of a vibrating sample magnetometer. It has been observed that for the duplex stainless steel, both filing as well as cold rolling promoted strain induced martensite. On the other hand, for the super duplex stainless steel, just filing promoted this transformation. In the comparing with duplex, the super duplex stainless steel austenite is more stable that is why is richer in nitrogen, so, the strain induced martensite formation is more difficult. The easier sigma phase precipitation during annealing as well in the super duplex stainless steel is due higher levels of chrome and molybdenum than the duplex stainless steel.

Les effets caloriques représentent la capacité d’un matériau à voir son entropie varier sous l’effet d’une sollicitation externe et peuvent être utilisés pour des systèmes de refroidissement à l'état solide en remplacement (ou complément) des dispositifs traditionnels à base de fluides frigorifiques. Dans cette thèse, nous avons cherché à étudier l'effet élastocalorique du caoutchouc naturel. Après une présentation des différents matériaux caloriques et du caoutchouc naturel, le chapitre 2 détaille la caractérisation élastocalorique du caoutchouc naturel, et les résultats sont interprétés à partir de la notion de cristallisation induite par la déformation. Le changement de température adiabatique élastocalorique et la variation d’entropie associée atteignent 9K et 50kJ.m-3.K-1, ce qui est très important comparé aux autres matériaux caloriques. L’effet élastocalorique étant maximum pour une déformation voisine de 4,5, une pré-déformation peut être appliquée pour éviter la zone moindre activité élastocalorique. La mesure directe de l’effet élastocalorique est ensuite comparée à une méthode indirecte déduite du facteur de Clapeyron, et les divergences sont discutées. Dans le chapitre 3, la contrainte et la température élastocalorique sont simulées par un modèle de Flory modifié sur la base de la cristallisation. Il est possible de prédire le comportement contrainte-déformation à différentes températures, ainsi que les variations de température élastocaloriques à température ambiante. Dans le chapitre 4, les effets de la fatigue sur l’effet élastocalorique du caoutchouc naturel sont ensuite étudiés. La résistance à la fatigue pour de grandes amplitudes de déformation est très faible (< 800 cycles). Trois régimes de déformation intermédiaire sont ensuite testés : 0-3, 2-5, et 4-7, et permet d’établir que le régime 2-5 est le plus performant (jusqu’à 100 000 cycles). Dans le dernier chapitre, un modèle de système régénératif de refroidissement à base de matériaux caloriques est développé afin d’établir des lignes directrices pour le choix des matériaux élastocaloriques / In this thesis, we aimed to study the eC effect of natural rubber (NR) and to prove its potential to act as an eC material primarily. The method for improving the eC effect efficiency and fatigue life of NR were also proposed. The eC effect of NR is characterized directly, and interpretation based on the theory of strain-induced crystallization/crystallite (SIC) is proposed. The eC adiabatic temperature change and isothermal entropy change of NR can be up to 9 K and 50 kJ.m-3.K-1 (56 J.kg-1.K-1), which are larger than most of caloric materials. Two coefficients, eC strain coefficient and eC stress coefficient , are defined for evaluating the eC performance at different strains, where is the specific entropy, is the engineering strain, is the temperature and is the stretching stress. It’s found that both coefficients are maximum for a strain around 4.5, indicating that the highest eC performance occurred at middle strain, which is attributed to the occurrence of SIC. To improve the eC performance, it is proposed to apply a pre-strain, so that the low strain regime where eC performance is low can be skipped. Moreover, the large needed deformation can be reduced by the pre-strain and thus the possibility of a compact cooling system designed based on NR is improved. The fatigue property of eC effect of NR is then investigated. The fatigue life at large deformation strain amplitudes (strain of 1-6) is about 800 cycles for the tested NR, which is too short to be used for a cooling system. Decreasing strain amplitude is necessary to extend fatigue life up to requirement of a cooling device. For the same small strain amplitude of 3, the fatigue property is compared at amorphous strain regime (strain of 0-3), onset strain of melting (strain of 2-5) and high strain of SIC (strain of 4-7). It’s found that a larger eC temperature change and a better fatigue property can be obtained at two SIC strain regimes (strain of 2-5 and 4-7) than amorphous strain regime. Especially, the fatigue property at the onset strain of melting (strain of 2-5) is better than that at high strain of SIC (strain of 4-7). A high-cycle fatigue was applied at the strain of 2-5 (most promising strain regime) up to 1.7×105 cycles. It was observed that there is no crack of the sample, as well as a degradation degree of 12% of the eC temperature change. Furthermore, the eC stress coefficient (4.4 K/MPa) at onset strain of melting is larger than that at high strain of SIC (1.6 K/MPa). As a result, the middle strain regime (onset strain regime of melting) can get a higher eC performance, larger temperature change, and better fatigue life, which should be chosen for eC cooling system.

Les effets électrocaloriques, qui se traduisent par une variation de température induite par une variation d’entropie ont été étudiés comme alternative aux systèmes de réfrigération utilisant un cycle de compression/détente. Le travail de thèse se focalise sur l’étude de l’effet élastocalorique dans le caoutchouc naturel et le terpolymère (P(VDF-TrFE-CTFE). En premier lieu, l’effet élastocalorique dans le caoutchouc naturel qui compte parmi les meilleurs candidats, a été évalué pour des cycles de déformation réalisés avec différentes valeurs d’allongement. Une variation de température de 4 °C a pu être observée. Il est usuel d’utiliser la relation déformation/contrainte en fonction de la variation de température pour évaluer l’effet élastocalorique. Il a été démontré que cette méthode ne peut pas être utilisée dans le cas du caoutchouc naturel et qu’elle doit être remplacée par la mesure de la variation de l’énergie mécanique en fonction de la température. Et dans ce cas, une variation linéaire entre ces deux dernières grandeurs a été observée. En réalisant un bilan d’énergie pendant l’essai, non seulement, le rendement énergétique a pu être évalué mais il a été aussi possible de prendre en compte l’effet Mullins et la cristallisation induite par la déformation pour le caoutchouc naturel. Dans un second temps, l’effet élastocalorique a été étudié sur le terpolymère (P(VDF-TrFE-CTFE), ce qui a permis de montrer qu’il était possible d’obtenir une variation de température de 2.1 °C sous réserver de pré-déformer le terpolymère à plus de 1050 % avant. Par comparaison avec d’autres matériaux présentant une bonne conversion élastocalorique, le fort potentiel de ce matériau a pu être mis en évidence. Enfin, il a été mis en évidence que la plus grande partie de l’énergie mécanique était bien convertie en énergie thermique. / Caloric effects (CEs), which are the phenomena that temperature variation is caused by entropy change, have been investigated for the novel system which might be able to replace conventional vapor compression refrigeration system. In the present thesis, the elastocaloric effect (ElCE) of natural rubber (NR) and terpolymer, poly(vinylidene fluoride-trifluoroethylene-chlorotrifluoroethylene) (P(VDF-TrFE-CTFE)), was focused. First of all, NR, which is an excellent candidate material for ElCE, was evaluated in cyclic deformation with different strain levels. It was found that NR exhibits temperature variation of around 4.0 °C. In general, the relation between stress/strain versus temperature variation is used to evaluate ElCE. The unsuitability of such evaluation method for NR was demonstrated. The evaluation method for ElCE which uses energy balance was then proposed. A linear relation between the temperature variation caused by ElCE and the applied mechanical energy by deformation was experimentally found. This fact verifies the suitability of the proposed method. Using the energy balance, besides, not only the conversion efficiency but also the influences of the Mullins effect and the strain-induced crystallization on the ElCE of NR were discussed. ElCE of P(VDF-TrFE-CTFE) was also evaluated in order to find out the potential of polymer. It was found that present terpolymer which is not one of the elastomers can also exhibit a large temperature variation, 2.1 °C, caused by ElCE if a large pre-stretch such as more than 1050 % is applied in advance. By comparison with other materials for ElCE, it was demonstrated that P(VDF-TrFE-CTFE) can be a high potential material for ElCE. It was also shown that P(VDF-TrFE-CTFE) converts most of the applied mechanical energy into the heat energy.

Os aços inoxidáveis dúplex possuem uma estrutura bifásica (ferrita e austenita) geralmente em frações aproximadamente iguais. Devido aos altos teores de Cr e Ni, esses aços apresentam alta resistência à corrosão e, por isso, são usados principalmente nas indústrias química, petroquímica e nuclear. Dependendo da sua composição química, os aços dúplex podem sofrer transformação martensítica induzida por deformação, com a transformação de austenita (&gamma;) em martensita (&alpha;\'). Essa transformação pode ser revertida mediante tratamento térmico. A austenita é paramagnética, enquanto que a ferrita e a martensita são ambas ferromagnéticas. O objetivo desse trabalho foi estudar a relação entre a estabilidade microestrutural e as propriedades magnéticas de um aço dúplex UNS S32304, o qual apresenta transformação martensítica induzida por deformação. Amostras desse aço com redução em espessura de 80% foram recozidas isotermicamente em várias temperaturas até 800ºC por 1 h e resfriadas em água. A partir de laços de histerese obtidos em temperatura ambiente foram obtidos os valores de magnetização de saturação (Ms) e campo coercivo (Hc) para essas amostras. Além das medidas magnéticas essas amostras foram caracterizadas via difração de raios X, testes de dureza, microscopias óptica (MO) e eletrônica de varredura (&epsilon;EV). Em relação ao &epsilon;EV foram utilizadas as técnicas de EBSD (do inglês \"eléctron backscatter diffraction\") e ECCI (do inglês \"electron channeling contrast imaging\"). Medidas de magnetização (in situ) em função da temperatura (até 1000ºC) também foram obtidas para o material deformado, a partir das quais foi determinada a temperatura de Curie (Tc) do mesmo. Uma simulação das fases presentes no material em função da temperatura foi obtida utilizando-se o software Thermo-Calc©. Para as amostras recozidas isotermicamente, a inspeção metalográfica mostrou que para 600-700ºC o material apresenta um aspecto fragmentado na microestrutura. Esse aspecto fragmentado é uma evidência da reversão da martensita em austenita. Para a amostra recozida em 700ºC precipitados foram encontrados principalmente na austenita, a qual parece estar recristalizada. Indícios de recristalização da ferrita também foram observados para essa amostra. De acordo com o Thermo-Calc© os precipitados observados são provavelmente do tipo M23C6 (M = Fe, Cr) e Cr2N. A dureza do material apresenta uma queda evidente a partir de 500ºC, relacionada aos fenômenos de recristalização do material e reversão da martensita. Tal como a dureza, Ms e Hc também decaem a partir de ~ 500ºC. A reversão da martensita em austenita e a decomposição da ferrita (&alpha;) contribuem para a diminuição de Ms. As medidas magnéticas in situ também forneceram evidências da transformação &alpha; &rarr; &gamma; + precipitados a partir do comportamento de Tc. Apesar do aspecto fragmentado da microestrutura e da precipitação, não foi observado um comportamento de \"pico\" em Hc em consequência da reversão da martensita em austenita. Isso indica que, no presente estudo, o fator que mais influenciou Hc foi a maior mobilidade das paredes de domínios magnéticos na fase ferrítica, em consequência dos efeitos de recuperação e recristalização. / Duplex stainless steels have a two-phase structure (ferrite and austenite) in approximately equal fractions. Due to their high Cr and Ni contents, these steels present a high corrosion resistance and, in consequence, they are mainly used in chemical, petrochemical and nuclear industries. Depending on their chemical composition, duplex steels may undergo strain induced martensite, with the austenite (&gamma;) transformation in martensite (&alpha;\'). This transformation can be reversed by annealing. The austenite is paramagnetic, whereas ferrite and martensite are both ferromagnetic. The goal of this work was to study the relationship between microstructural stability and magnetic properties of a UNS S32304 duplex steel, which presents strain induced martensite. Samples of this steel with 80% thickness reduction were isotermally annealed at several temperatures up to 800ºC for 1 h and water-cooled. From hysteresis loops taken at room temperature, both saturation magnetization (Ms) and coercive field (Hc) were obtained for these samples. In addition to magnetic measurements these samples were characterized using X-ray diffraction, hardness testing, optical (OM) and scanning electron (SEM) microscopies. Regarding SEM were used both EBSD (eléctron backscatter diffraction) and ECCI (electron channeling contrast imaging) techniques. In situ magnetization measurements in function of temperature (up to 1000ºC) were also performed for the deformed material, from which was determined its Curie temperature (Tc). A simulation of the phases present in the material as a function of temperature was performed using the Thermo-Calc© software. For the isothermally annealed samples, metallographic analysis showed that for 600-700ºC the material presents a fragmented microstructure. Such fragmentation is an evidence of the martensite-to-austenite reversion. For the sample annealed at 700ºC precipitates were found mainly in the austenitic phase, which appears to be recrystallized. Evidences of recrystallization were also found for the ferritic phase in the same sample. According to Thermo-Calc© the observed precipitates are probably M23C6 (M = Fe,Cr) and Cr2N. Material\'s hardness present an evident drop for temperatures higher than 500ºC, due to both recrystallization and martensite-to-austenite reversion phenomena. Like hardness, both Ms and Hc also drop in temperatures higher than 500ºC. The martensite-toaustenite reversion and the ferrite decomposition contribute to Ms decreasing. From the Tc behavior, the in situ magnetic measurements also provided evidences of the transformation &alpha; &rarr; &gamma; + precipitates. In spite of microstructure fragmentation and precipitation, it was not observed a \"pick\" effect in Hc behavior as a consequence of the martensite-to-austenite reversion. This indicates that, in the present study, the factor that most influenced Hc was the higher mobility of magnetic domain walls in the ferritic phase, due to both recovery and recrystallization effects.

Le caoutchouc naturel présente une très bonne résistance à la propagation de fissure. Cette particularité est généralement attribuée dans la littérature à sa capacité à cristalliser sous déformation. A ce jour, l'essentiel des travaux dans ce domaine porte sur des échantillons réticulés par une vulcanisation dite efficace, c’est-à-dire dont les noeuds de réticulation sont principalement composés de ponts monosulfures. Pour certaines applications, et parce qu’elle est réputée conduire à de meilleure propriétés en fatigue, il est intéressant d'utiliser une vulcanisation dite conventionnelle. Le matériau est alors composé majoritairement de ponts polysulfures. Le phénomène de cristallisation sous déformation, exacerbée en pointe de fissure en raison d'une amplification du champ de déformation, semble être l'un des mécanismes responsables de l'accroissement de la durée de vie en fatigue du matériau vulcanisé de manière efficace. Cependant, sur matériaux vulcanisés de manière conventionnelle et thermo-oxydés (77°C), cette corrélation doit être confirmée en raison d’une évolution importante de l'architecture du réseau élastomère pendant son vieillissement pouvant en effet avoir un impact important sur la capacité du matériau à cristalliser sous déformation. Cette étude se propose donc de caractériser l’évolution de réseaux élastomères vulcanisés de manières conventionnelles pendant leur vieillissement thermo-oxydant, puis d’évaluer leur résistance à la propagation de fissure à différents niveaux de déformation macroscopique. Des analyses in situ WAXS sous rayonnement synchrotron en fond de fissure permettent alors de relier ces résultats à l'évolution de l’aptitude de ces matériaux à cristalliser sous déformation. / Natural rubber (NR) is largely used in the tire industry due to its excellent mechanical properties, e.g. its very good resistance to fatigue crack growth at high strain. It is generally accepted that this outstanding behavior is related to its ability to crystallize under strain. Such phenomenon, so called SIC, strongly depends on parameters like temperature, strain rate as well as the architecture of the rubber network. The microstructure of this network is formed during the crosslinking process and depends on the vulcanization system, i.e. “Efficient” or “Conventional”. The former vulcanization recipe consists in the formation of short or monosulfide bridges in the elastomer network whereas the latter (necessary to ensure a good adhesion between metallic and rubber parts in a tire) will mainly create longer polysulfide bridges. During its life, the tire will be submitted to a slow aerobic ageing which will cause structural modifications of the initial network and therefore an evolution of the rubber ability to crystallize under strain and to resist against crack propagation. In general, the structural modifications are caused by complex chemical mechanisms, highly sensitive to temperature, leading to chain scission and chain crosslinking. They can also involve sulfur bridge reorganization when NR is conventionally vulcanized. Nevertheless, most of the literature on NR ageing has been performed on efficiently cross-linked NR, and in thermal conditions which are much too severe to be representative of the material ageing in tire applications. Within this frame, our objective is to study this material when it is aged at 77 °C in air. Such parameters have been identified as capable of reproducing more realistically and over a reasonable duration, the ageing of rubber in some use conditions. After characterization of the evolution of the aged materials microstructure, their crack propagation resistance will be studied at 0.01 Hz for different values of macroscopic deformations. Time resolved Wide Angle X-ray scattering (WAXS) measure-ments, carried out at room temperature, will then provide information on the crystallization process around the crack tip. Based on these results, the relation between the network evolution during ageing, the fatigue properties and the ability to strain crystal-lize in such conditions will be established in this work.

Inconel 738 (IN 738), like other precipitation-hardened nickel-base superalloys that contain a substantial amount of Al and Ti, is very difficult to weld due to its high susceptibility to heat-affected zone (HAZ) cracking during conventional fusion welding processes. The cause of this cracking, which is usually intergranular in nature, has been attributed to the liquation of various phases in the alloy, subsequent wetting of the grain boundaries by the liquid and decohesion along one of the solid-liquid interfaces due to on-cooling tensile stresses. To address the problem of liquation cracking in weldments, recent developments in welding research have resulted in supposedly exclusive solid-state friction joining processes, such as linear friction welding (LFW), for joining crack susceptible structural alloys. The objective of this work was therefore to investigate the weldability of the difficult-to-weld IN 738 superalloy by LFW and to analyze the resulting microstructural changes in the alloy due to the welding process. LFW was performed on Linear Friction Welding Process Development System (PDS) at the Aerospace Manufacturing Technology Centre of the Institute for Aerospace Research, National Research Council (NRC) of Canada. In order to study and decouple the effect of non-equilibrium thermal cycle and imposed compressive stress during the joining, physical simulation of the LFW process was performed by using Gleeble 1500-D Thermo-Mechanical Simulation System at the University of Manitoba. Detailed microstructural study of welded and Gleeble-simulated materials was carried out. Correlation between the simulated microstructure and that of the weldments was obtained, in that, a significant grain boundary liquation was observed in both the simulated specimens and actual weldments due to non-equilibrium reaction of second phase particles, including the strengthening gamma prime phase. These results show that in contrast to the general assumption of LFW being an exclusively solid-state joining process, intergranular liquation, caused by non-equilibrium phase reaction(s), occurred during the process. However, despite a significant occurrence of liquation in the alloy, no HAZ cracking was observed. Nevertheless, the result showed that crack-free welding by linear friction welding is not due to preclusion of grain boundary liquation as has been commonly assumed and reported. Instead, resistance to cracking can be related to the counter-crack-formation effect of the imposed strain and to a concept observed and reported for the first time in this work, which is strain-induced rapid solidification. Furthermore, microstructural evolution during joining cannot be understood without considering the concept of non-equilibrium liquation reaction and strain-induced rapid solidification of the metastable liquid, which are carefully elucidated in this thesis.

Inconel 738 (IN 738), like other precipitation-hardened nickel-base superalloys that contain a substantial amount of Al and Ti, is very difficult to weld due to its high susceptibility to heat-affected zone (HAZ) cracking during conventional fusion welding processes. The cause of this cracking, which is usually intergranular in nature, has been attributed to the liquation of various phases in the alloy, subsequent wetting of the grain boundaries by the liquid and decohesion along one of the solid-liquid interfaces due to on-cooling tensile stresses. To address the problem of liquation cracking in weldments, recent developments in welding research have resulted in supposedly exclusive solid-state friction joining processes, such as linear friction welding (LFW), for joining crack susceptible structural alloys. The objective of this work was therefore to investigate the weldability of the difficult-to-weld IN 738 superalloy by LFW and to analyze the resulting microstructural changes in the alloy due to the welding process. LFW was performed on Linear Friction Welding Process Development System (PDS) at the Aerospace Manufacturing Technology Centre of the Institute for Aerospace Research, National Research Council (NRC) of Canada. In order to study and decouple the effect of non-equilibrium thermal cycle and imposed compressive stress during the joining, physical simulation of the LFW process was performed by using Gleeble 1500-D Thermo-Mechanical Simulation System at the University of Manitoba. Detailed microstructural study of welded and Gleeble-simulated materials was carried out. Correlation between the simulated microstructure and that of the weldments was obtained, in that, a significant grain boundary liquation was observed in both the simulated specimens and actual weldments due to non-equilibrium reaction of second phase particles, including the strengthening gamma prime phase. These results show that in contrast to the general assumption of LFW being an exclusively solid-state joining process, intergranular liquation, caused by non-equilibrium phase reaction(s), occurred during the process. However, despite a significant occurrence of liquation in the alloy, no HAZ cracking was observed. Nevertheless, the result showed that crack-free welding by linear friction welding is not due to preclusion of grain boundary liquation as has been commonly assumed and reported. Instead, resistance to cracking can be related to the counter-crack-formation effect of the imposed strain and to a concept observed and reported for the first time in this work, which is strain-induced rapid solidification. Furthermore, microstructural evolution during joining cannot be understood without considering the concept of non-equilibrium liquation reaction and strain-induced rapid solidification of the metastable liquid, which are carefully elucidated in this thesis.

In an attempt to improve the quality in finite element analysis of thermoformed components, a method for predicting the thickness distribution is presented. The strain induced anisotropic material behaviour in the amorphous polymers of concern is also taken into account in the method. The method comprises of obtaining raw material data from experiments, followed by a simulation of the vacuum thermoforming process where hyperelastic material behaviour is assumed. The theory of hyperelasticity that was applied was based on the Ogden model and implemented in the FE-software LS-DYNA. Material behaviour from thermoformed prototypes is examined by experiments and implemented together with the mapped results from the thermoforming simulation in a succeeding FE-model. For the latter, the three-parameter Barlat model was suggested, giving the possibility to account for anisotropic material behaviour based on an initial plastic strain.

A presente tese de doutorado avalia a Emissão Magnética Espontânea (SME), fenômeno descrito recentemente no grupo de pesquisa do professor Hélio Goldenstein e que já demonstrou ser uma ferramenta promissora para monitorar a transformação martensítica nos aços. O SME é utilizado para monitorar a velocidade de propagação de plaquetas de martensita; identificar a temperatura em que ocorre o inicio da primeira transformação (Ms); observar martensita induzida por deformação ou tensões durante a deformação de aços contendo austenita metaestável e também estudar a transformação isotérmica da austenita em martensita. A duração dos picos individuais magnéticos emitidos no início (Ms) e no final (Mf) da transformação martensítica durante o resfriamento contínuo foram medidos e comparados com os tamanhos médios das placas de martensita, obtidos através da metalografia quantitativa. O SME, um fenômeno mais sensível do que outras medidas globais como a resistividade, dilatometria, Ruído Magnético de Barkhausen (MBN), etc., não exige um volume crítico de transformação para ser detectado, e é capaz de detectar sinais de placas individuais crescentes. / This PhD thesis evaluates the Spontaneous Magnetic Emission (SME), a phenomenon recently described in the research group of Professor Hélio Goldenstein that has proved to be a promising tool to monitor the Martensitic Transformation in steels. The SME is used to monitor the speed of propagation of martensite plates; to identify the temperature at which the first transformation occurs (Ms); to observe martensite straininduced or stresses during deformation of steels containing metastable austenite and also study the isothermal transformation of austenite in martensite. The duration of individual magnetic peaks issued at the start (Ms) and in the end (Mf) of the martensitic transformation during the continous cooling were measured and compared with the average size of the martensite plates obtained by quantitative metalography. The SME, a phenomenon more sensitive than other global measures such as resistivity, dilatometry, Magnetic Barkhausen Noise (MBN), etc., does not require a critical volume of transformation to be detected, and is able to detect signs of growing individual plates.