O objetivo deste estudo foi comparar a eficácia anestésica da articaína 4%, da lidocaína 2% e da mepivacaína 2%, todas associadas à epinefrina 1:100.000, durante pulpectomia em pacientes com pulpite irreversível em molares mandibulares. Sessenta e seis voluntários do Setor de Urgência da Faculdade de Odontologia da Universidade de São Paulo receberam, aleatoriamente, 3,6ml de um dos anestésicos locais para o bloqueio convencional do nervo alveolar inferior (NAI). No caso de falha do bloqueio, foram administrados 3,6ml da mesma solução como injeção complementar no ligamento periodontal. O sinal subjetivo de anestesia do lábio, a presença de anestesia pulpar e ausência de dor durante a pulpectomia foram avaliados, respectivamente, por indagação ao paciente, por meio do aparelho estimulador pulpar elétrico (pulp tester) e por uma escala analógica verbal. A análise estatística foi realizada por meio dos testes Qui-quadrado, Kruskal Wallis e Razão de Verossimilhancas. Todos os pacientes reportaram anestesia no lábio após o bloqueio do NAI. A mepivacaína apresentou valores superiores (68,2%) para a anestesia pulpar após o bloqueio do NAI e a lidocaína (90%) após a injeção no ligamento periodontal. A mepivacaína apresentou valores superiores para a analgesia (72,7%) após o bloqueio no NAI e a lidocaína (90%) após a injeção no ligamento periodontal. Após a falha do bloqueio do NAI, a dor na câmara pulpar foi a mais frequente e após a falha da injeção no ligamento periodontal, a dor no canal. No entanto, essas diferenças não foram estatisticamente significantes. Portanto as três soluções anestésicas locais se comportam de forma semelhante e não apresentam efetivo controle da dor no tratamento da pulpite irreversível em molares mandibulares. / The aim of this study was to compare the anesthetic efficacy of 4% articaine, lidocaine 2% and 2% mepivacaine, all associated with epinephrine 1:100,000 during pulpectomy in patients with irreversible pulpitis in mandibular molars. Sixty-six volunteers Sector Emergency Faculty of Dentistry, Universidade de São Paulo randomly received 3.6 ml of a local anesthetic to block conventional inferior alveolar nerve. In case of failure of the lock, were administered 3.6 ml of the same solution as in the periodontal ligament injection complement. The signal subjective lip anesthesia, the presence of pulpal anesthesia and no pain during pulpectomy were evaluated respectively by questioning the patient, via the stimulating device electrical pulp (pulp tester) and a verbal analogue scale. Statistical analysis was performed using the chi-square test, Kruskal Wallis and likelihood ratio. All patients reported lip anesthesia after blockade of the inferior alveolar nerve. The mepivacaine showed higher values (68.2%) for pulpal anesthesia after blockade of the inferior alveolar nerve and lidocaine (90%) after injection in the periodontal ligament. The mepivacaine showed higher values for analgesia (72.7%) after blocking the inferior alveolar nerve and lidocaine (90%) after injection in the periodontal ligament. After the failure of the blockade of the inferior alveolar nerve, the pain in the pulp chamber was the most frequent and after the failure of the periodontal ligament injection, pain in the channel. However, these differences were not statistically significant. Therefore, the three local anesthetic solutions behave similarly and did not exhibit effective pain management in treating irreversible pulpitis in mandibular molars.

Este trabalho busca descrever sistemas irreversíveis (aqueles que não obedecem ao balanceamento detalhado) usando o formalismo mecânico-estatístico que tem como base a dinâmica estocástica. Nossos principais objetivos são: (i) a investigação do comportamento crítico e das possíveis classes de universalidade em sistemas irreversíveis; (ii) a modelagem da dinâmica de propagação de epidemias. Primeiramente investigamos o modelo suscetível-infectado-recuperado (SIR) estocástico e espacialmente estruturado. Nesse modelo, os indivíduos são divididos em três classes: suscetível (S), infectado devido ao contato com um vizinho infectado, e um individuo infectado pode recuperar-se espontaneamente. Este modelo exibe transição de fase em que a doença se espalha e uma fase em que não há espalhamento da doença. Tratando cada par suscetível-infectado como uma conexão através da qual pode haver propagação da epidemia, mostramos que é possível estabelecer uma conexão entre o modelo SIR e o modelo de percolação. Assim, pudemos utilizar métodos da teoria de percolação usual para determinar o limiar de espalhamento epidêmico. Por meio de aproximações de campo médio dinâmico, simulações computacionais de Monte Carlo estacionárias e simulações dependentes do tempo, determinamos o ponto critico e o comportamento critico desse modelo. Ademais, propomos dois modelos para descrever um processo epidêmico de transmissão vetorial. Consideramos duas populações interagentes uma formada por vetores e a outra por hospedeiros. Os vetores podem ser suscetíveis (S) ou infectados (I), enquanto os estados permitidos para os hospedeiros são os mesmos do modelo SIR. O processo de transmissão da doença ocorre devido ao contato local de um hospedeiro (vetor) suscetível com um vetor (hospedeiro) infectado. Determinamos o limiar de infecção, o tamanho da epidemia e mostramos que ambos os modelos exibem transições de fase de segunda ordem e que pertencem à classe de universalidade da percolação dinâmica isotrópica. / This study aims to describe irreversible systems (those that do not obey detailed balance) using a statistical mechanics formalism based on stochastic dynamics. Our main objectives are: (i) to investigate the critical behavior and the possible universality classes in irreversible systems; (ii) and modeling the dynamics of epidemic spreading. First we investigate the stochastic and spatially structured susceptible-infected-recovered model (SIR). In this model, individuals are divided into three classes: susceptible (S), infected (I) and recovered (R). A susceptible individual may become infected due to contact with an infected neighbor, and an infected individual may recover spontaneously. This model exhibits a phase transition between a phase in which the epidemic spreads and a phase where there is no spreading of the disease. Treating each susceptible-infected pair as a connection through which there may be epidemic spreading, we show that it is possible to establish a connection between the SIR model and the percolation model. Thus we are able to use methods of the theory of standard percolation for determining the epidemic spreading. By means of dynamic mean-field approximations and stationary and time-dependent computational Monte Carlo simulations, we determine the critical point and critical behavior of this model. In addition, we propose two models to describe the vector transmitted epidemic process. We consider two interacting populations, one formed by vectors and other by hosts. The vectors may be susceptible (S) or infected (I), where the states allowed for the hosts are the same as those in the SIR model. Transmission of the disease occurs due to contact between a local host (vector) and a susceptible vector (host) infected. We determine the threshold of infection, the size of the epidemic, and show that both models exhibit second order phase transitions that belong to the universality class of dynamic isotropic percolation.

O objetivo deste estudo foi comparar a eficácia anestésica de um volume de 1,8mL do cloridrato de articaína 4% com epinefrina 1:100.000 com um volume de 3,6mL do mesmo anestésico local no bloqueio convencional do nervo alveolar inferior (BNAI) e na injeção complementar no ligamento periodontal em pacientes com pulpite irreversível de molares mandibulares. Noventa pacientes do Setor de Urgência da Faculdade de Odontologia da Universidade de São Paulo receberam, aleatoriamente, 1 (1,8mL) ou 2 (3,6mL) tubetes da solução anestésica no BNAI. No caso de falha do BNAI, foram administrados os mesmos volumes pré-selecionados, aleatoriamente, na injeção complementar do ligamento periodontal. O sinal subjetivo de anestesia do lábio, a presença de anestesia pulpar e ausência de dor durante o procedimento de pulpectomia foram avaliados, respectivamente, por indagação ao paciente, por meio do aparelho estimulador pulpar elétrico e por uma escala analógica verbal. A análise estatística foi realizada por meio dos testes Qui-quadrado, Kruskal Wallis e Razão de Verossimilhancas. Após o BNAI, todos os pacientes reportaram anestesia no lábio. O volume de 1,8mL de articaína apresentou 27% de anestesia pulpar e o volume de 3,6mL apresentou 42%. A analgesia para o grupo de 1,8mL foi de 64% e para o volume de 3,6mL foi 73% após o BNAI, porém, essas diferenças não foram estatisticamente significantes. Na ocorrência da falha do BNAI, 64% dos pacientes sentiram dor na câmara pulpar, 32% em dentina e 4% no canal. Após a injeção no ligamento periodontal, o grupo de 1,8mL apresentou 75% de anestesia pulpar e o grupo de 3,6mL apresentou 42%. Em relação à analgesia durante o procedimento de pulpectomia, 31% dos pacientes do grupo de 1,8mL e 25% dos pacientes do grupo de 3,6mL apresentaram dor após a injeção no ligamento periodontal, porém, essas diferenças não foram estatisticamente significantes. Na ocorrência da falha da injeção do ligamento periodontal, 62% dos pacientes sentiram dor na câmara pulpar, 25% em dentina e 13% no canal. O aumento do volume de 1,8mL para 3,6mL da solução de articaína 4% com epinefrina 1:100.000 no BNAI e na injeção complementar no ligamento periodontal não aumentou significativamente a taxa de sucesso da anestesia pulpar e da analgesia. Portanto, os dois volumes anestésicos se comportam de forma semelhante, e não são efetivos no controle da dor durante o tratamento da pulpite irreversível de molares mandibulares.. / The aim of this study was to compare the anesthetic efficacy of a 1.8mL volume of 4% articaine hydrochloride with epinephrine 1:100,000 with a volume of 3.6mL of the same anesthetic in the inferior alveolar nerve block (BNAI) during pulpetcomy procedure in patients with irreversible pulpitis in mandibular molars. Ninety patients from the Emergency Department of the School of Dentistry of the University of São Paulo received randomly 1 (1.8mL) or 2 tubes (3.6mL) of the anesthetic solution in the BNAI. In the case of failure of the BNAI, the same pre-selected volumes were administered in the complementary injection of the periodontal ligament. The subjective signal of lip anesthesia, the presence of pulp anesthesia and absence of pain during pulpectomy were evaluated, respectively, by patient inquiry, through the electrical pulp stimulator and by an analogical verbal scale. Statistical analysis was performed using the Chi-square, Kruskal Wallis and Reason of Verossimilhancas tests. After the BNAI, all patients reported anesthesia on the lip. The volume of 1.8mL of articaine presented 27% of pulpal anesthesia and the volume of 3.6mL presented 42%. Analgesia for the 1.8mL group was 64% and for the volume of 3.6mL it was 73% after the BNAI, but these differences were not statistically significant. In the occurrence of BNAI failure, 64% of the patients felt pain in the pulp chamber, 32% in dentin and 4% in the root. After injection into the periodontal ligament, the 1.8mL group had 75% of pulpal anesthesia and the 3.6mL group had 42%. Regarding analgesia during the pulpectomy procedure, 31% of the patients in the 1.8mL group and 25% of the patients in the 3.6mL group had pain after injection into the periodontal ligament, but these differences were not statistically significant. In the occurrence of failure of the periodontal ligament injection, 62% of the patients felt pain in the pulp chamber, 25% in dentin and 13% in the root. Increasing the volume from 1.8mL to 3.6mL of the 4% articaine solution with 1: 100,000 epinephrine in the BNAI and in the complementary injection in the periodontal ligament did not significantly increase the success rate of pulpal anesthesia and analgesia. Therefore, both anesthetic volumes behave similarly, and are not effective in controlling pain in the treatment of irreversible mandibular molar pulpitis.

Made available in DSpace on 2014-08-20T14:30:13Z (GMT). No. of bitstreams: 1 Tese_Gloria_Beatriz_Azevedo_Cubas.pdf: 1660996 bytes, checksum: 8bb44113acc263526d63d6f5e7e47e2b (MD5) Previous issue date: 2012-12-07 / The aims of this randomised controlled trial were to evaluate the influence of a preprocedural 0.12% chlorhexidine mouthrinse on the microbial contamination of dental impressions, subsequently disinfected with sodium hypoclorite or water (control). The second objective was to evaluate if aqueous solution of 0.12% chlorhexidine mixed with irreversible hydrocolloid powder would decrease microbial contamination of dental impressions. Forty subjects underwent maxillary dental impressions with irreversible hydrocolloid and were randomly divided into two groups (n=20) according to the preprocedural mouthrinse (0.12% chlorhexidine or placebo).The dental impressions were then divided into two subgroups and disinfected with sodium hypochlorite or water (control). In the second part, 20 subjects underwent maxillary dental impressions with irreversible hydrocolloid and randomly assigned into two groups according to the mixed agent used (0.12% chlorhexidine or water). Saliva and alginate samples were assessed for microbiological counts of total micro-organisms, total streptococci and Candida species. Surface roughness of the impressions and dimensional stability of the casts were also evaluated. Chlorhexidine preprocedural mouthrinse significantly reduced (p<0.05) microbial contamination. Small but significant alterations were produced on dimensional stability and surface quality when sodium hypochlorite was used as disinfectant after the impression (p=0.005). The results also showed that the use of aqueous 0.12% chlorhexidine mixed with the powder of irreversible hydrocolloid decreased the percentage of total micro-organisms and total Streptococci counts (p<0.001), without producing alterations in surface quality and dimensional stability. It can be concluded that 0.12% chlorhexidine used as preprocedural mouthrinse, or mixed with irreversible hydrocolloid powder, are effective methods in reducing microbial contamination of alginate impressions, without causing damage to the physical and mechanical properties of the material / Os objetivos deste ensaio clínico randomizado foram (I) avaliar a influência de um bochecho, com solução de clorexidina 0,12% e placebo, previamente a realização de moldagens bucais, posteriormente desinfetadas com hipoclorito de sódio ou água (controle) na contaminação de moldagens de alginato; e (II) avaliar se solução aquosa de clorexidina 0,12% misturada com o pó de hidrocolóide irreversível reduziria a contaminação microbiana de moldagens bucais.Quarenta voluntários foram submetidos a moldagens bucais com hidrocolóide irreversível e foram divididos de forma randomizada em 2 grupos (n=20) de acordo com a solução bucal (0,12% clorexidina ou placebo). As moldagens bucais foram divididas em 2 grupos e desinfetadas com hipoclorito de sódio ou água (controle). Na segunda parte do ensaio clinico, 20 voluntários foram submetidos a moldagens bucais com hidrocolóide irreversível e foram divididos em 2 grupos de acordo com a solução usada no preparo do alginato (0,12% clorexidina ou água).Amostras de saliva e alginato foram avaliados quanto a contagem microbiana de microrganismos totais, estreptococos totais e espécies de Candida. Rugosidade de superfície e estabilidade dimensional de modelos de gesso também foram avaliados. Bochecho com solução bucal de clorexidina reduziu de forma significativamente (p<0,005) a contaminação microbiana. Pequenas mais significativas alterações dimensionais e de rugosidade de superfície foram produzidas quando hipoclorito de sódio foi utilizado como agente desinfetante pós-moldagem (p=0,005). Os resultados também demonstraram que o uso de solução de clorexidina 0,12% misturada ao pó de hidrocolóide irreversível reduziu a percentagem de microrganismos totais e estreptococos (p<0.001), sem causar alterações de rugosidade de superfície e estabilidade dimensional. Pode ser concluído que o uso de solução de clorexidina 0,12% usada como bochecho bucal ou misturada ao pó de hidrocolóide irreversível são métodos eficientes na redução microbiana de moldagens de alginato, sem causarem danos as propriedades físicas e mecânicas do material

L’objectif de ce travail est de proposer de nouvelles stratégies de commande non linéaire pour la stabilisation des Réacteurs Parfaitement Agités Continus (RPAC). Pour cela, nous utilisons d’une part, l’approche thermodynamique entropique. Plus précisément, nous utilisons la notion de disponibilité thermodynamique et les propriétés de la thermodynamique irréversible pour définir une fonction de Lyapunov utilisable pour la stabilisation du système en boucle fermée. Nous proposons aussi une fonction disponibilité réduite afin d’obtenir des lois stabilisantes plus performantes en terme de sollicitation des actionneurs. D’autre part, nous proposons une extension du formalisme (pseudo) hamiltonien à ports dissipatifs aux réacteurs chimiques ouverts. Nous montrons que l’Hamiltonien est lié à l’enthalpie libre de Gibbs dans le cas isotherme et à l’ectropie (opposée de l’entropie) dans le cas non isotherme. Par ce formalisme, la dissipation du système représente la production irréversible d’entropie due à la réaction chimique. Nous appliquons ensuite les techniques de commande passive (modelage de l’énergie) pour la synthèse de lois de commande en choisissant la disponibilité thermodynamique comme fonction hamiltonienne à modeler en boucle fermée. Finalement, nous montrons que les commandes synthétisées par l’approche thermodynamique entropique et la formulation pseudo-hamiltonienne sont, dans certains cas, équivalentes. Certaines propriétés relatives à la stabilisation et l’admissibilité des commandes sont aussi considérées. Les développements théoriques sont mis en oeuvre sur des exemples différents de RPAC : un réacteur académique et l’hydrolyse par catalyse acide de l’oxirane-méthanol en glycérine. / The goal of this thesis is to propose new nonlinear control strategies for the stabilization of perfectly Continuous Stirred Tank Reactors (CSTR). To achieve this goal, we use on the one hand, the entropic thermodynamic approach. More precisely, we use the thermodynamic availability concept and the properties of irreversible thermodynamics to define a Lyapunov function candidate for the stabilization of the closed loop system. We also propose a reduced availability function to design more efficient feedback laws in term of control variable solicitations. On the other hand, we propose an extension of the (pseudo) Hamiltonian formalism associated to dissipative systems to open chemical reactors. We show that the Hamiltonian is linked to the Gibbs free enthalpy in the isothermal case and to ectropy (opposed to entropy) in the non isothermal case. By this formalism, the dissipation of the system represents the irreversible entropy production due to chemical reaction. The Interconnection and Damping Assignment-Passivity Based Control (IDA-PBC) approach is then applied to synthesize feedback laws by choosing the thermodynamic availability as desired closed loop hamiltonian storage function. Finally, we show that feedback laws synthetized by the entropic thermodynamic approach and the pseudo-hamiltonian formulation are equivalent in some cases. Some stabilization properties and the control input admissibility are also considered. Theoretical developments are illustrated on some different CSTR examples : an academic case study and the acid catalyzed hydration of oxirane-methanol to glycerol.

L’objet de cette thèse est la modélisation et la commande par approche thermodynamique des réacteurs catalytiques triphasiques en mode continu et en mode discontinu. Ce type de réacteur consiste en un système fortement non linéaire, multivariable et siège de réactions exothermiques. Nous utilisons les concepts de la thermodynamique irréversible pour la synthèse de lois de commande stabilisante pour ces deux types de réacteurs chimiques. En effet, la stricte concavité de la fonction d’entropie nous a permis de définir une fonction de stockage qui sert de fonction de Lyapunov candidate : la disponibilité thermodynamique. Nous utilisons cette fonction de disponibilité thermodynamique pour la synthèse de lois de commande stabilisante d’un mini-réacteur catalytique triphasique intensifié continu. Une stratégie de contrôle à deux couches (optimisation et contrôle) est utilisée pour contrôler la température et la concentration du produit à la sortie du réacteur en présence de perturbations à l’entrée du réacteur. Les performances du contrôleur mis en place sont comparées en simulation à celles d’un régulateur PI. Dans certains cas, l’utilisation de la fonction de disponibilité thermodynamique s’avère problématique. Une autre étude effectuée sur cette fonction nous permet de déterminer une nouvelle fonction de Lyapunov : la disponibilité thermique. Nous utilisons par la suite la fonction de disponibilité thermique pour la synthèse de lois de commande stabilisante d’un réacteur catalytique triphasique semi-fermé. Un observateur grand gain est utilisé pour estimer la vitesse de réaction à partir des mesures de la température du milieu réactionnel. Cette estimation est injectée ensuite dans le calcul de la loi de commande mise en place. La robustesse du schéma de contrôle est testée en simulation face à des incertitudes de modélisation, des perturbations et des bruits de mesure. / The goal of this thesis is the modeling and thermodynamic based control of three-phase catalytic reactor working in continuous or discontinuous modes. These types of reactors are highly nonlinear, multivariable and exothermal processes. We use the concepts of irreversible thermodynamics for the synthesis of stabilizing control laws for these two types of chemical reactors. Indeed, the strict concavity of the entropy function has allowed us to define a storage function used as a candidate Lyapunov function: the availability function. We use this availability function for the synthesis of control laws for stabilizing a three-phase catalytic continuous intensified mini-reactor. A control strategy with two layers (optimization and control) is used to control the temperature and concentration of the product at the outlet of the reactor in the presence of disturbances. The performances of the controller are compared by simulation to those of a PI controller. In some cases, the use of the availability function may cause some problems. A new Lyapunov candidate function is then derived from the original availability function: the thermal availability. We use this thermal availability for the synthesis of stabilizing control laws for a three-phase catalytic fed-batch reactor. A high gain observer is used to estimate the chemical reaction rate from the measurements of the temperature. This estimate is then used for the control law implementation. The robustness of the control scheme is tested in simulation against modelling uncertainties, disturbances and noise measurements.

Le transport des bruts paraffiniques, et tout particulièrement leur remise en écoulement après un arrêt, dans de longues conduites sous-marines soumises à de basses températures, peut être difficile du fait de l'augmentation de leur viscosité. Le comportement rhéologique d´une huile paraffinique modèle, possédant des propriétés macroscopiques d'écoulement analogues à celles des bruts paraffiniques, est d'abord analysé en utilisant la vélocimétrie par imagerie par résonance magnétique associée à des mesures de contrainte de cisaillement au sein d'une géométrie Couette. Nous montrons que lors d'un écoulement forcé à température constante le matériau subit une déstructuration irréversible qui dépend de l'intensité du cisaillement. Ainsi la contrainte apparente critique permettant l'écoulement du matériau dépend de l'histoire thermique et d'écoulement subie par le matériau. Nous étudions ensuite le comportement rhéologique complet de deux bruts réels à partir de différents types de tests rhéométriques (fluages, redémarrage, régime permanent, changement brusque de vitesse) pour différentes histoires d'écoulement, notamment pendant la période de refroidissement. Le comportement détaillé du matériau en régime transitoire ainsi observé peut alors être modélisé. De plus les variations du seuil de contrainte en fonction de l'histoire thermique et de l'écoulement sont aussi décrites, ce qui nous donne le champ de contrainte seuil dans la conduite à l'état initial. Le modèle dans son ensemble est finalement implémenté dans un code de calcul pour simuler le redémarrage de l´écoulement d'un brut paraffinique dans une conduite réelle / Transporting waxy crude oils through long pipelines at low temperatures may be challenging, particularly its flow restart after a pipeline shut-in, due to its viscosity increase. The rheological behavior of a model waxy oil with macroscopic flow properties analogous to waxy crude oils is first analyzed using Magnetic Resonance Imaging velocimetry associated to stress measurements in a Couette geometry. While flowing at constant temperature, major irreversible structure break depending on shear intensity are observed. Thus, the critical apparent shear stress beyond which the material flows depends on the thermal and flow histories of the oil. Next, the rheological behavior of two waxy crude oils is studied using rheometrical tests (creep tests, flow restarts, abrupt changes of shear rate and steady flow) after different flow histories, notably during the cooling process. Then, those experimentally observed trends are modeled. Additionally, a comprehensive study of the yield stress in function of flow and temperature histories is presented. It provides an approach for describing the yield stress field inside the pipeline at the flow restart moment. Finally, the entire rheological model is implemented in the computational code for simulating waxy crude oils flow restart of a real scale pipeline

Cette thèse se positionne au croisement de la théorie de la décision en univers incertain et de la théorie des choix d’investissements irréversibles (options réelles). Elle poursuit trois objectifs principaux :1. Tout d’abord, elle s’inscrit dans un courant de recherche dynamique, notamment en économie et en finance, qui vise à modéliser l’impact de l’ambigüité à laquelle des décideurs sont parfois confrontés lorsqu’ils contemplent des choix aux conséquences irréversibles. 2. Ensuite, elle met l’accent sur la persistance de fortes controverses théoriques portant sur les fondements axiomatiques des modèles de décision face à l’ambigüité. Aussi, nous proposons d’utiliser certaines propriétés des modèles non linéaires pour aborder sous un angle original la représentation de l’ambigüité et des préférences des individus face à celle-ci. En particulier, nous suggérons de ne pas restreindre a priori la nature des préférences individuelles face à l’ambigüité. Pour cela, nous adoptons les fondements de l’approche de Choquet, à savoir tout particulièrement l’emploi de capacités (probabilités non additives) pour pondérer les différentes alternatives ambigües. Tout en proposant ce processus stochastique ambigu, dit Choquet-Brownien, nous soulignons les conditions de l’inévitable arbitrage entre réalisme des hypothèses et souplesse d’utilisation du modèle. D’un point de vue axiomatique, une attention particulière est portée au respect de la cohérence dynamique.3. Enfin, cette thèse vise à encourager une prise en considération plus ambitieuse des sources d’incertitude dans le cadre des options réelles. Alors qu’ils sont présentés comme des outils privilégiés pour affronter le risque, les modèles d’options réelles ont certainement beaucoup à gagner à s’enrichir par la prise en compte également de l’ambigüité. En effet, alors que le risque est largement discuté dans la littérature des options réelles, l’impact de l’ambigüité est très largement ignoré. / The need to elaborate innovative methods to analyze risk and uncertainty has become increasingly obvious over the last decades, especially due the growing perception of the multiplicity of social and economical issues characterized by the weight of uncertainty (natural disasters, ecological risk, financial crises…).This thesis is at the crossroad between decision theory under uncertainty and the irreversible investment theory (real options). Consequently, the main goal of this thesis is three-fold: 1. First, it contributes to the dynamic stream of literature in economics and finance that models the impact of ambiguity that individuals may often face and/or perceive when contemplating irreversible choices.2. Next, this thesis emphasizes that even with the plethora of decision models already dealing with uncertainty, elaborating sound axiomatic foundations largely remains an open question. This leads us to recommending the use of non linear models (such as multiple-priors, Choquet expected utility, robust control, smooth ambiguity), which in turn raises many challenging theoretical and practical obstacles. We explore original ways of addressing some of these issues and suggest the construction of ambiguous stochastic processes in a Choquet expected utility framework (that are called Choquet-Brownian motions): ambiguity preferences are thereby directly embedded into the trajectory of some random variables that may drive a decision, such as the expected cash flows of an investment project or its exit value.3. Finally, this thesis also aims specifically at encouraging the enrichment of real option models. It is striking that only the impact of risk has been widely discussed by the real option theory so far, while the specific impact of ambiguity has been largely ignored. Considering that the real option theory is directly concerned with sources of flexibility, irreversibility and uncertainty in general, ambiguity represents a promising expansion.

This dissertation thesis describes a phenomenon called electroporation. It is about its theoretical aspects as well as about modeling of processes in the tissue during electroporation. Further, it describes the technical design of two developed unique experimental generators of DC and AC pulses for electroporation purposes. It also includes a description of experiments which were done using discussed generators.

The settling of spherical and cylindrical particles in a linearly stratified fluid is investigated using experiments. The double-tank method is used to generate a linear stratification with a red colored dye homogeneously mixed in the heavy water tank. As a result of feeding the stratification using dyed heavy water, the concentration of dye varies with depth in the experiment tank. A powerful back-light and a digital camera are used to record the events. Assuming the concentration of dye is directly proportional to density of fluid, Beer-Lambert's law is used to generate a calibration between intensity of the light measured by the camera and density of the fluid. Using this calibration, density is evaluated in all the images captured. In the parameter space of this study, the spheres have three different wake patterns. The area of fluid disturbed by a suspension of spheres increases with <i>Re</i> and <i>Fr</i>. As a result, the amount of energy available for the mixing and the irreversible change of total potential energy in the system increases with <i>Re</i>, <i>Fr</i> and number of particles. Cylinders drag volumes of light fluid to larger depths in their wake than spheres and shed the light fluid in the form of vortices. This results in lower volumes of fluid perturbed by the cylinders. However, as the light fluid is dragged to larger depths, the amount of energy generated for mixing and the change in total potential energy of the system is higher. Spheres are thus more efficient in disturbing volumes of fluid but cylinders are more efficient in causing irreversible changes to the state of the system.