This thesis deals with a high temperature testing of fracture toughness and studies the size effect on measured values using miniature size CT specimen. Two types of specimen geometry were manufactured from P91 steel, the standard size and the quarter size specimen. J-R curves were obtained in the temperature range from 23°C to 600°C. No specimen size effect was observed at room temperature tests. The realized experiments together with fractography analysis demonstrated the drop of toughness at 400°C caused by the effect of dynamic strain aging.

De récents travaux consacrés à l’étude des propriétés des matériaux aux petites échelles ont souligné des différences exceptionnelles dans le comportement mécanique des nano-objets par rapport aux matériaux massifs. Dans le cas du silicium, une transition fragile-ductile à température ambiante a été observée lorsque la taille des échantillons est réduite. Cependant, les défauts et les mécanismes à l’origine de ce changement de comportement n’ont pas été clairement identifiés. Ce travail repose sur l’étude post mortem de nanopiliers déformés, en utilisant différentes techniques de microscopie électronique. Les nanopiliers étudiés ont été préparés par gravure plasma et déformés en compression à température ambiante. Les résultats obtenus durant cette thèse, confirment la différence de comportement des nano-objets par rapport au matériau massif. Par ailleurs, une grande variété de défauts produits lors de la compression a été observée. L’orientation cristallographique de l’axe de sollicitation semble avoir un impact important sur les mécanismes à l’origine du comportement ductile observé. La comparaison entre images HRTEM expérimentales et simulées témoigne de la propagation simultanée de dislocations partielles et parfaites dans les plans {111}. De plus, des événements plastiques ont également été observés dans des plans {115}. Divers mécanismes de déformation possibles impliqués lors de la compression des piliers sont décrits à partir des observations microscopiques. Un modèle tenant compte de l’influence sur la mobilité des dislocations des interactions entre systèmes de glissement est proposé afin d’expliquer la transition fragile-ductile observé aux petites échelles / Several recent works devoted to the study of the properties of materials at small scales have revealed exceptional differences in the mechanical behavior of nano-objects as compared to bulk material. In the case of silicon, a brittle-ductile transition at room temperature has been observed when the sample size decreases. However, the extended defects and mechanisms behind this behavioral change have not been clearly identified. This work is based on the post mortem study of deformed nanopillars, using different electron microscopy techniques. The studied nano-pillars of 100 nm in diameter were prepared by plasma etching (RIE) and deformed in compression at room temperature. The results obtained during this thesis confirm the difference in the behavior of nano-objects compared to bulk material. Moreover, a large variety of defects produced during plastic deformation has been observed. The crystallographic orientation of the deformation axis seems to have a significant impact on the mechanisms behind the observed ductile behavior. The comparison between experimental and simulated HRTEM images notably evidences the simultaneous propagation of partial and perfect dislocations in {111} planes. In addition, unexpected plastic events have also been observed in {115} planes. Various possible deformation mechanisms involved during the nano-compression of the pillars are described, based on the microscopic observations. Finally, a model considering the influence of interactions between various activated systems on the mobility of dislocations is proposed to explain the brittle-ductile transition observed at small scales in silicon

Ces travaux s’attachent à décrire et quantifier les mécanismes de ruptures des compositestissés 3D sous chargement de traction quasi-statique et à mettre en place une méthode de simulationnumérique adaptée et robuste, pouvant à terme être appliquée en bureau d’études.Dans cette optique, une étude expérimentale a été menée afin de quantifier la propagation defissures dans ces matériaux. Celle-ci a permis de mettre en place un scenario de rupture, entirant parti de la multi-instrumentation des essais. L’étude a également été effectuée sur deséprouvettes de géométries et de tailles variées et a mis en évidence d’importantes variations dutaux de restitution d’énergie avec les conditions d’essai. Un formalisme d’analyse et de modélisationintroduisant des longueurs internes a ensuite été présenté et adapté aux mécanismes derupture des composites tissés 3D. Ce formalisme est étayé par la recherche des mécanismes àl’aide de l’analyse des faciès de rupture. Les longueurs introduites ont ainsi été mises en relationavec les paramètres du tissage. Une méthode d’identification des paramètres a été proposée etles conséquences de ce comportement sur le dimensionnement de pièces composites discutées.Enfin, le transfert de ces résultats a été effectué vers des simulations numériques robustes. Desméthodes de régularisation des modèles d’endommagement continu ont été présentées et évaluéesà l’aune de leur capacité à assurer, d’une part, la robustesse des résultats et, d’autre part,la bonne retranscription des effets d’échelle expérimentaux. La prise en compte de ces considérationsnumériques et physiques nous a amené à proposé un modèle d’endommagement Non-Local.Une méthode d’identification des paramètres et de la longueur interne à partir des données expérimentalesa été proposée. / This work aims to describe and quantify the failure mechanisms of 3D woven composites underquasi-static tensile loading and to implement an adapted and robust numerical simulationmethod, that can be applied in industry. To this end, an experimental study was carried out toquantify the propagation of cracks in these materials. Thus, a crack propagation scenario wasestablished, thanks to the multi-instrumentation used during the tests. The experimental campaignwas carried out on specimens of various geometries and sizes and highlighted significantvariations in the fracture toughness with the test conditions. A modelisation framework introducinginternal lengths was then presented and adapted to 3D woven composites. This frameworkis supported by the identification of the failure mechanisms subsequent to the analysis of thecrack profile. The introduced lengths were thus related to the weaving parameters. A method foridentifying the parameters was proposed and the consequences of this behaviour on the designof the composite parts discussed. Finally, these results were transferred to robust numerical simulations.Regularisation methods of continuous damage models were presented and evaluatedin terms of their ability to ensure, on the one hand, the robustness of the results and, on theother hand, the correct transcription of experimental size effects. Taking into account these numericaland physical considerations led us to propose a Non-Local damage model. A method foridentifying the parameters and the internal length on experimental data was proposed.

We developed a multiscale approach to explore the effective thermal conductivity of polycrystalline graphene sheets. By performing equilibrium molecular dynamics (EMD) simulations, the grain size effect on the thermal conductivity of ultra-fine grained polycrystalline graphene sheets is investigated. Our results reveal that the ultra-fine grained graphene structures have thermal conductivity one order of magnitude smaller than that of pristine graphene. Based on the information provided by the EMD simulations, we constructed finite element models of polycrystalline graphene sheets to probe the thermal conductivity of samples with larger grain sizes. Using the developed multiscale approach, we also investigated the effects of grain size distribution and thermal conductivity of grains on the effective thermal conductivity of polycrystalline graphene. The proposed multiscale approach on the basis of molecular dynamics and finite element methods could be used to evaluate the effective thermal conductivity of polycrystalline graphene and other 2D structures.

Detta arbete syftar till att undersöka relationen mellan företagsstorlek och dess aktieavkastning,annars känt som storlekseffekten, inom halvledarsektorn. Vi använder oss av två portföljer bestående av de tio största och tio minsta halvledarbolagen och görutfallstestet under perioden 2004–2015. Tre olika allokeringsstrategier tillämpas: equal weight, meanvariance och equal risk contribution samt tre olika ombalanseringsperioder. Vårt resultat visar på ett negativt samband mellan företagsstorlek och riskjusterad avkastning oavsettallokeringsstrategi. Resultaten tyder på att effekten inte är en proxy för fundamentala skillnader ellerberor på en felspecificering av β. / This thesis aims to examine the relationship between firm size and stock return, otherwise known asthe size effect, within the semiconductor industry. We construct two portfolios each comprising the ten largest and smallest semiconductor companiesand conduct a back test between 2004-2015. We examine three allocation strategies: equal weight,mean variance, and equal risk contribution along three difference rebalancing periods. Our results show a negative relationship between firm size and risk adjusted return regardless ofallocation strategy. The results also show that size effect is not a proxy for fundamental differencesnor a misspecification of β.

This thesis was most concerned with the mechanical response to irradiation of two in-house produced oxide dispersion strengthened (ODS) steels and two non-ODS coun- terparts. The steels, manufactured by Dr. M. J. Gorley (University of Oxford), were me- chanically alloyed from gas-atomised Fe-14Cr-3W-0.2Ti, with the addition of 0.25Y₂O₃ powder in the case of the ODS variants. The powders were hot isostatic pressed at consolidation temperatures of 950 °C and 1150 °C. The four steels were designated 14WT 950 (non-ODS), 14YWT 950 (ODS), 14WT 1150 (non-ODS) and 14YWT 1150 (ODS), and were used in the as-produced condition. Initially, the macroscale elastic modulus and yield stress were determined using a four-point flexure test, employing digital image correlation (DIC) as a strain gauge. The microcantilever size eects were then characterised, and it was determined that the yield stress signicantly diverged from macroscale values at microcantilever beam depths of < 4.5 μm. Using knowledge of this, the in-house produced alloys were irradiated with 2 MeV protons at the Surrey Ion Beam Centre (University of Surrey, UK) to a displacement damage of ∼ 0.02 dpa and 0.2 dpa (Bragg peak). This was to produce a deep irradiated layer for the fabrication of large microcantilevers with reduced size effects. The cross-sectional surface of the irradiated layer was then exposed and inclined linear arrays of 250 nm deep indents were placed across the damage prole. 14WT 1150 (non-ODS) revealed a clear proton damage prole in plots of hardness against irradiation depth, 14WT 950 (non-ODS) also showed modest hardening in the region of the Bragg peak. No appreciable hardening was observed in either 14YWT specimens, attributed to the fine dispersion of nanoscale oxides providing a high number density of defect sink sites. However, a large bimodal variation in hardness was measured in both ODS variants. This was investigated using EBSD and EDX, and was determined to be caused by a pronounced heterogeneity of the microstructure. While Hall-Petch strengthening and changes in the local chemistry had some effect on the measured hardness, the most likely cause of the large variation in local hardness was heterogeneity in the nanoscale oxide population. Microcantilevers were fabricated out of the irradiated layer cross-section in 14WT 1150 and 14YWT 1150. Larger microcantilevers, with ∼ 5 μm beam depth, were placed with their beam centre at ∼ 0.026 dpa. Smaller microcantilevers, with ∼ 1.5 μm beam depth, were placed with their beam centre at the Bragg peak, 0.2 dpa. Both the large and the small microcantilevers fabricated in 14WT 1150 (non-ODS) displayed significant irradiation hardening. In the ODS variant, 14YWT 1150, irradiation hardening appeared to be reduced. The work in this thesis successfully showed that it was possible to extract a close approximation of the macroscale yield stress from shallow irradiated layers, providing that the irradiation condition is carefully chosen in response to known size dependent behaviour. This thesis also investigated the size dependent behaviour of microcantilevers using a lengthscale dependent crystal plasticity UMAT, developed by Dunne et al. and implemented within ABAQUS 6.14-2 commercially available nite element software. The simulation of the GND density evolution with increasing plastic strain allowed their contribution to the microcantilever size effect, through mobile dislocation pinning, to be determined. This novel approach to modelling size effects in three dimensional finite element microcantilever models demonstrated that while it was possible to simulate a lengthscale-dependent response in finite element microcantilever models, the constitutive equation for the plastic velocity gradient needs to be more physically based in order the match the experimentally derived results; for example, a lengthscale-dependent term relating to the dislocation source density of the material. Although the apparent reduction of irradiation hardening in ODS in-house produced alloys showed great promise, these alloys also displayed a large amount of scatter in measured hardness and yield stress, attributed to the pronounced heterogeneity in the microstructure. Alloys with such signicant microstructural heterogeneity are not suitable for engineering or commercial use.

1997年爆發亞洲金融風暴，隔年（1998年）起公司發生財務危機事件層出不窮，1998年至2005年間最為嚴重；2007年全球金融海嘯至今，投資人擔心買到地雷股，對於投資股票市場仍採觀望態度。在經過層層把關的財報背後究竟隱藏多少危機？這些危機難道是不可預測的嗎？其實，公司爆發財務危機並非一夕之間產生的問題，就如同人類的慢性病不是一天造成的，是長期忽略身體健康警訊造成的結果，事出必有因，因此許多學者便開始探究財務危機背後的成因，試圖找出一些指標供投資人作為投資前之考量因素。 本研究主要之目的在於探究財務危機之預測指標，分別探討Beta、公司規模、淨值市價比以及前一年平均報酬（負的動能效果）是否可作為財務危機之預測指標。本研究之樣本公司為1983年至2007年之台灣上市公司，利用Altman提出之Z-score模型將公司區分為危機公司以及正常公司，再將樣本公司依Beta、公司規模、淨值市價比以及前一年平均報酬分別分組，探討這些變數是否可作為財務危機之預測指標。實證結果指出Beta及淨值市價比無法作為財務危機之預測指標，但公司規模及前一年平均報酬（負的動能效果）可以作為財務危機之預測指標。 / With the Asian financial crisis breaking out in 1997, many companies began to suffer financial distress in the following year, and the situations were getting even worse during 1998 and 2005. Faced the new waves of financial tsunami across the world starting from 2007, the investor, therefore, have been adopting a wait-and-see attitudes towards the stock market, fearing of being hit by the “tank stocks”。How many financial problems hidden behind the carefully prepared financial statements? Are they unpredictable? As a matter of fact, just like the human chronic diseases which actually caused by long-term ignorance of health warning, corporate financial distress never happens suddenly. Thus a number of scholars are dedicated to study the reasons for financial problems, attempting to figure out certain indicators capable of being prior reference for investment decision-making. This paper aims to study the predictors of financial distress. Beta, firm size, book-to-market ratio and average monthly prior-year return (negative momentum effect) are to be considered respectively to determine their possibilities of being predictors. The sample companies discussed in this paper are chosen among the listed companies during 1983 and 2007 in Taiwan. They are grouped into two categories of crisis company and normal company by using the Z-score model developed by Altman. Then the sample companies are divided in terms of Beta, firm size, book-to-market ratio and average monthly prior-year return so as to trace these variables’ likelihood to predict bankruptcy. It eventually turns out that firm size and average monthly prior-year return could serve as predictors of financial distress, other than Beta and book-to-market ratio.

Ce travail concerne des expériences mécaniques, des analyses numériques et des modélisations analytiques de la rupture cohésives (Mode I), vis-à-vis de l’étude du comportement mis en évidence par la courbe de Résistance (Courbe-R) et l’effet d’échelle de structures entaillées en bois massif. Des expériences de fissuration sont combinées à des analyses numériques pour déterminer les propriétés de rupture au moyen d’une procédure appelée Théorie de la Mécanique de la Rupture Linéaire Élastique équivalente (TMRLE), basée sur la complaisance de la structure. La courbe-R, obtenue à partir des expériences, selon une méthode de correction du poids propre, montre l’existence d’un domaine endommagé (Zone de Processus de Rupture) de taille non négligeable se développant en fond de fissure. Dans des conditions de fissuration stationnaire, ce domaine atteint une taille critique, et l’énergie nécessaire pour faire propager la fissure avec ce domaine endommagé (par unité de surface de rupture), reste constante. Le taux de libération de l’énergie de fissuration ainsi attendu, joue un rôle important en Mécanique de la Rupture, car il est possible simuler le comportement quasi-fragile du matériau en combinaison avec les autres propriétés de cohésion. La loi d’effet d’échelle de Bažant, utilisée pour prévoir l’influence de la taille sur la contrainte nominale, est estimée à partir de la réunion de deux comportements asymptotiques réalisés sur de petites tailles (Analyse limite ou RdM) et des grandes tailles. Une procédure analytique est présentée pour évaluer le comportement asymptotique additionnel exhibé par la contrainte nominale dans le régime intermédiaire, de façon plus exacte. Une validation numérique est présentée, et l’information expérimentale vient confirmer ce comportement asymptotique. / This work concerns the mechanical testing, numerical analysis and modelling of cohesive fracture (Mode I) on the purpose to study the Resistance-curve behaviour and the size effect in wooden notched structures. The mechanical testing is combined with the numerical analysis to evaluate fracture properties by means of an equivalent LEFM approach based on the structure compliance. The Resistance-curve being revealed from the experiments, by means of a self-weight compensation method, correction puts into evidence that a non-negligible damaged domain (Fracture Process Zone) is under development in the crack front during the loading process. This being the case, among other fracture parameters issued from the Resistance-curve, the critical (asymptotic) energy release rate is determined, turning possible to use it in combination with other cohesive crack properties in the crack modelling (in Mode I). Thus, for a given geometry it is possible to monitor the critical dimension being revealed by the Fracture Process Zone (FPZ) during the crack propagation. The well known Bažant’s size effect law provides the scaling of the nominal strength through the asymptotic matching performed both on the small (Strength Theory) and on the large (LEFM) structure sizes. An analytical procedure is proposed to determine an additional asymptotic regime in the intermediate size range through a more accurate manner. Numerical validations of the proposed procedure are made and experimental data is presented revealing the scaling of the nominal strength through an envelop of values.

Il est de plus en plus admis que l’effet d’échelle doit être pris en compte dans la conception des structures de Génie Civil. Pour le béton, ce problème est complexe car celui-ci ne possède pas d’adoucissement plastique, et sa rupture est due à la fissuration caractérisée par une grande zone de microfissuration (fracture process zone) qui dépend de la taille du granulat max d .Cette fissuration passe par un adoucissement sous la forme de microfissures et de glissement interparticules. Expérimentalement, l’effet d’échelle sur le béton est très souvent étudié à l’aide des corps d’épreuves homothétiques entaillés où l’on cherche à relier la résistance nominale ( oN )estimée à partir de la charge de rupture en flexion à une dimension caractéristique D. Ceci conduit à une diminution du ratio dmax/D avec l’augmentation de la taille de la structure. Parmi les objectifs de cette thèse est d’étudier expérimentalement l’impact de l’hétérogénéité ( dmax/D)supposé comme facteur fondamental de l’effet d’échelle. Trois coupures granulaires ont été testées sur trois tailles de poutres différentes en suivant le processus de fissuration par émission acoustique et la technique de corrélation d’images. Celles-ci permettent de suivre l’ouverture des fissures et identifient assez clairement la FPZ. Les résultats mettent en évidence une grande influence de la taille du granulat sur le comportement à la rupture du béton. Il existe une relation directe entre les paramètres de l’effet d’échelle obtenus par la loi de Bazant et la taille du granulat( dmax ). Le traitement des résultats d’une même taille avec différents granulométries dans le même diagramme conduit à la même loi d’effet d’échelle structurelle classique avec une valeur de transition identique. La modélisation du comportement mécanique est effectuée par la méthode d’éléments discrets (DEM). Le modèle de contact linéaire ne s’avère pas adéquat pour le mortier et le béton où le rapport compression / traction est très élevé. De ce fait, Il a été modifié pour prendre en compte la contribution des moments inter-granulaires. Les paramètres micromécaniques sont déterminés par des essais classiques avec une analyse inverse en utilisant l’algorithme de Levenberg-Marquardt. Les résultats montrent que cette approche est capable de reproduire le comportement à la fissuration locale du béton et de reproduire l’effet d’échelle et celui des granulats. Ensuite, un modèle d'adoucissement est développé afin de mieux reproduire la réponse post pic et le processus de fissuration. / It is now commonly understood that in the design of civil engineering structures, size effect must be taken into consideration. For concrete, this problem is complex because it does not exhibit plastic softening. The failure of concrete is generally preceded by propagation of cracks, characterized by alarge microcracking zone (fracture process zone or FPZ) which is proportional to the maximum aggregate size ( dmax ). This fracture process is accompanied by strain-softening in the form of microcracking and fractional slip.Experimentally, size effect in concrete is commonly studied by using geometrically similar notched beams where thenominal strength ( oN ) obtained from the bending failure loadis related to the characteristic dimension (D). This leads to adecrease in the ratio of dmax/D with an increase in the size of the structure. One of the objective of this thesis is to study experimentally the effect of heterogeneity ( dmax/D) size. This ratio is recognized as a fundamental factor causing the size effect. Three aggregate grading segments were tested on three different sizes of beams and the cracking process was investigated by acoustic emission and the image correlation technique. These methods make it possible to trace the crack.openings and identify distinctively the FPZ. The results demonstrate a significant influence of the aggregate size on the fracture behaviour of concrete. There is a direct relationship between the size effect parameters obtained by Bazant's law and maximum aggregate size ( dmax ). The results obtained from the specimen having the same size but made of concretes with different aggregate sizes produced the same classical size effect with identical transitional between LEFM and strength based laws. The mechanical behaviour is modelled by the Discrete Element Method (DEM). However, the linear contact model inserted in DEM is not suitable to satisfy the materials like mortar and concrete with high unconfined compressive strength to tensile strength ratio. As a result, the model is modified to take into account the contribution of interparticle moments. The micromechanical parameters are determined by conventional tests with inverse analysis using the Levenberg-Marquardt algorithm. The results showed that this approach is able to reproduce the local cracking behaviour of concrete as well as classical size effect and aggregate size effect. Then, a softening model is developed to better reproduce the post-peak response and the cracking process.

Le travail présenté dans ce document vise à mettre en évidence et à comprendre l'effet de la taille nanométrique des renforts sur les propriétés des nanocomposites avec une approche expérimentale. Des nanocomposites de PMMA et particules de silice (15nm, 25nm, 60nm, 150nm et 500nm) de fractions volumiques 2 0/0, 40/0 et 6 0/0 ont été fabriqués. Des analyses multi-échelles (MET et DRX-WAXS) ont montré que les paramètres caractéristiques de la microstructure des nanocomposites varient avec la taille des nanoparticules. En effet, la diminution de la taille des nanoparticules à fraction volumique constante a entrainé une diminution de la distance intermoléculaire. Cette diminution a induit une densification de la matrice et une réduction de la mobilité des chaînes de la matrice. Des essais mécaniques (traction, DMA) ont montré que les modules de Young (E) et de conservation (E') des nanocomposites augmentent avec la diminution de la taille des nanoparticules à fraction volumique constante. Et que l'augmentation de E' est conservée avec l'augmentation de la température. Une augmentation des températures de transition vitreuse (Tg) et de dégradation (Td) a également été observée avec les essais DSC, DMA et ATG. Le modèle de la borne inférieure d'Hashin-Shtrikman étendue aux nanocomposites à renforts sphériques proposé par Brisard a été utilisé. La modélisation des modules élastiques des nanocomposites a montré que pour reproduire les données expérimentales, il faut que d'une part que les modules surfaciques caractérisant l'interface soient dépendants de la taille des nanoparticules. Et d'autre part, tenir compte de l'état de dispersion des nanoparticules. / The work presented in this paper aims to highlight and to understand the size effect of nano-reinforcements on nanocomposite properties With an experimental approach. Nanocomposites of PMMA and silica particles With different sizes (15nm, 25nm, 60nm, 150nm and 500nm) and volume fractions (20/0, 4 0/0 and 60/0) were manufactured. Multiscale analysis (MET and DRX-WAXS) have shown that the characteristic parameters of the microstructure of nanocomposites vary With the size of the nanoparticles. Indeed, the decrease in the size of nanoparticles at a given volume fraction implies a decrease of the intermolecular distance. This decrease has induced a densification of the matrix and a decrease of the matrix chain mobility. Mechanical tests (tensile, DMA) have shown that the young (E) and the conservation (E') moduli of the nanocomposites increase With the decrease in the size of the nanoparticles With a constant volume fraction. And the increase of E l is kept when temperature growing. An increase in glass transition (Tg) and degradation temperature (Td) was also observed With the DSC, DMA and ATG tests. Experimental elastic properties of the nanocomposites were used to assess the relevance of size effect micromechanical models, particularly the Hashin-Shtrikman bounds With interface effects proposed by Brisard. The modeling has shown that to reproduce the experimental elastic moduli of nanocomposites, the elastic coefficients of the interface must be dependents on particle sizes. And the state of dispersion of particles must be taken into account.