This thesis project investigates the state-of-the-art in traffic management "Green Light Optimal Speed Advisory (GLOSA)" for vehicles in a smart city. GLOSA utilizes infrastructure and vehicles communication through using current signal plan settings and updated vehicular information in order to influence the intersection approach speeds. The project involves traffic microscopic simulations for a mixed traffic environment of conventional and automated vehicles (AVs) both connected to the intersection control and guided by a speed advisory traffic management system. Among the project goals is to assess the effects on traffic performance when human drivers comply to the speed advice. The GLOSA management approach is accessed for its potential to improve traffic efficiency in a full market penetration of connected AVs with absolute compliance. The project also aims to determine the possible outcome resulting from enhancing the AVs capabilities such as implementing short time headways between vehicles in the future.  The best traffic performance results achieved by operating GLOSA goes for connected AVs with the lowest simulated time headway (0.3 sec). The waiting time reduction reaches 95% and trip delay lessens to 88 %.

Les simulations hydrodynamiques pour la physique de haute densité d'énergie ainsi que pour la fusion par confinement inertiel exigent une description détaillée de flux d'énergie. Le mécanisme principal est le transport électronique, qui peut être un phénoméne non local qui doit être décrit avec des modèles de Fokker-Planck, stationnaires et simplifiés dans les codes hydrodynamiques à grande échelle. Mon travail thèse est consacré au développement d'un nouveau modèle de transport non local basé sur l'utilisation d'une méthode de fermeture entropique pour la résolution des premiers moments de l'équation de Fokker-Planck agrémentée d'un opérateur de collision dédié. Une telle fermeture permet une bonne résolution des fortes anisotropies de la fonction de distribution électronique dans les régimes où le développement d'instabilités électrostatiques à petite échelle le requiert. Ce modèle aux moments (M1) est comparé avec succès au modèle de Schurtz, Nicolaï et Busquet (SNB), référent dans le domaine du transport électronique non local. Ce modèle, basé sur l'hypothèse d'une faible anisotropie de la fonction de distribution sous-jacente induisant une relation de fermeture polynomiale (P1), utilise un opérateur de collision simplifié dont nous avons proposé une amélioration. Après avoir considéré plusieurs configurations typiques de transport de chaleur, nous avons montré que le modèle M1 ultidimensionnel peut prendre naturellement en compte des effets d'un plasmas magnétisés sur le transport électronique. De plus, ce modèle permet de calculer des fonctions de distribution utiles aux études cinétiques comme la stabilité du plasma dans la zone de transport. Nous confirmons avec notre modèle que le transport d'énergie électronique peut fortement modifier l'amortissement des ondes de Langmuir et des ondes acoustiques ; contrairement aux modèles non locaux simplifiés, M1 décrit les modifications de la fonction de distribution et l'amortissement des ondes du plasma. La structure du modèle permet également de prendre en compte naturellement des champs magnétiques autogénérés, qui jouent un rôle crucial dans des simulations multidimensionnelles. Ces champs magnétiques pourraient également être étudiés pour concentrer l'énergie dans les schémas d'ignition. Enfin, nous montrons que le modèle M1 reproduit les résultats de la théorie locale élaborée par Braginskii pour tous les niveau de magnétisation et propose de nouveaux résultats pour le régime non local. Ce travail constitue une première validation de l'utilisation des fermetures entropiques, dans les régimes de faibles anisotropies, qui va s'ajouter aux tests dans les régimes fortement anisotropes. / Hydrodynamic simulations in high-energy-density physics and inertial con nement fusion require a detailed description of energy uxes. The leading mechanism is the electron transport, which can be a nonlocal phenomenon that needs to be described with quasistationary and simplified Fokker-Planck models in large scale hydrodynamic codes. My thesis is dedicated to the development of a new nonlocal transport model based on a fast-moving-particles collision operator and on a first moment Fokker-Planck equation, simplified with an entropic closure relation. Such a closure enables a better description of the electron distribution function in the limit of high anisotropies, where small scale electrostatic instabilities could be excited. This new model, so called M1, is successfully compared with the well known nonlocal electron transport model proposed by Schurtz, Nicolaï and Busquet, using different collision operators, and with the reduced Fokker-Planck model, based on a small-anisotropies polynomial closure relation (P1). Several typical configurations of heat transport are considered. We show that the M1 entropic model may operate in two and three dimensions and is able to account for electron transport modifications in external magnetic fields. Moreover, our model enables to compute realistic electron distribution functions, which can be used for kinetic studies, as for the plasma stability in the transport zone. It is demonstrated that the electron energy transport may strongly modify damping of Langmuir and ion acoustic waves, while the simplified nonlocal transport models are not able to describe accurately the modifications of the distribution function and plasma wave damping. The structure of the M1 model allows to naturally take into account self-generated magnetic fields, which play a crucial role in multidimensional simulations. Moreover, magnetic fields could also be used for the focusing of energetic particles in alternative ignition schemes. The M1 model reproduces the results of the local transport theory in plasma, developed by Braginskii, in a broad range of degrees of magnetization and predicts new results in the nonlocal regime. This work constitutes a first validation of the entropic closure assumption in the weakly-anisotropic regime. It can be added to the existing tests, in the strongly-anisotropic regimes.

Selecionamos consecutivamente 63 pacientes com doença mista do tecido conectivo (DMTC) (Kasukawa, 87) para determinar a relevância do padrão SD. Ter uma capilaroscopia periungueal (CPU) até cinco anos antes do início do estudo foi o principal critério de inclusão. Na entrada, avaliamos o envolvimento de órgãos e os auto-anticorpos. A idade média e o tempo de doença foram 45,3 + 10 e 8,45 + 5,42 anos, respectivamente. O padrão SD foi observado em 41 pacientes na entrada (65%) e em 45 na CPU prévia (71,5%), p = 0,20. Dez pacientes (16%) alteraram a CPU, 7 normalizaram e 3 desenvolveram padrão SD. O tempo de doença, número e freqüência de órgãos envolvidos foram semelhantes em pacientes com e sem padrão SD. Em contraste, a análise de cada parâmetro do padrão SD mostrou uma freqüência significativamente menor de áreas avasculares (AA) moderadas/graves na entrada, comparada com a CPU anterior (26,5 e 53%, p = 0,013). Além disto, 76% dos pacientes com doença intersticial pulmonar (TCAR) tiveram AA na entrada, enquanto apenas 24% dos pacientes com esta alteração não apresentavam este achado à CPU (p = 0,017). Adicionalmente, reduzida densidade capilar foi freqüentemente observada em pacientes submetidos à terapia imunossupressora, quando comparados com o grupo sem este tratamento (66,7 e 33,3%, p = 0,001). A CPU na DMTC é um processo dinâmico e a análise de cada parâmetro do padrão SD parece ser um bom indicador de doença intersticial pulmonar e gravidade de doença. / For determining the clinical relevance of SD-pattern in MCTD, sixty-three MCTD patients (Kasukawa´s criteria) were consecutively selected. The main inclusion criterion was availability of previous nailfold capillaroscopy (NC) 5 years before inclusion. At entry, organ involvement and autoantibody evaluation were performed. The mean age and disease duration were 45.3 + 10 and 8.45 + 5.42 years, respectively. SD-pattern was observed in 41 patients at entry (65%) and in 45 at previous NC (71.5%), p = 0.20. Ten patients (16%) changed NC, 7 normalized, and 3 developed SD-pattern. Disease duration, number and frequency of organ involvement were similar in patients with and without SD-pattern. In contrast, analysis of each SD-pattern parameter revealed a significant lower frequency of moderate/severe avascular areas (AA) at entry compared to previous examination (26.5 vs. 53%, p = 0.013). Moreover, 76% of patients with interstitial lung disease (HRCT) had AA at entry, whereas only 24% of patients with this alteration did not have this NC finding (p = 0.017). Furthermore, reduced capillary density was frequently observed in patients taking immunosuppressive therapy than those without (66.7 vs. 33.3%, p = 0.001). NC in MCTD is a dynamic process and analysis of each SD-pattern parameter seems to be a good indicator of lung involvement and disease severity

Les travaux présentés visent à améliorer les modèles de physique nucléaireutilisés dans l’évaluation des sections efficaces neutroniques de fission. Le résultat deces travaux donne les clefs pour une percée significative dans ce domaine et a permisd’étendre fortement les capacités du code d’évaluation CONRAD. Les sections partiellesétant naturellement corrélées entre-elles pour respecter la valeur de la section totale, cesaméliorations bénéficient à l’ensemble des sections partielles. Un cadre solide pour lamodélisation des processus concurrent à la fission a dû être établi sur le modèle du codede référence TALYS. Après s’être assuré de la fiabilité et de la cohérence du cadre, lesinvestigations spécifiques concernant la fission ont pu être réalisées. Les perspectivesd’applications offertes par les modèles macro-microscopiques FRDM et FRLDM ont étéanalysées. Ces modèles ont été implémentés et validés sur des données expérimentaleset des benchmarks. Afin d’obtenir des temps de calcul compatibles avec les besoins del’évaluation, des méthodes numériques sophistiquées ont été sélectionnées et une partiedes calculs a été portée sur GPU. Ces modèles macro-microscopiques peuvent être utiliséspour construire des surfaces d’énergie potentielle qui sont à leur tour traitées afin d’obtenirdes barrières de fission à une dimension, puis des coefficients de transmission fission. Cesderniers sont alors utilisés dans le cadre de modélisation des sections efficaces moyennesdu domaine statistique sur la base d’un modèle Hauser-Feshbach. Les résultats de cetteapproche seront présentés sur le cas du 239Pu(n,f). / The work presented here aims to improve models used in the fission crosssectionevaluation. The results give insights for a significant breakthrough in this fieldand yielded large extensions of the evaluation code CONRAD. Partial cross sections areinherently strongly correlated together as of the competition of the related reactions mustyield the total cross section. Therefore improving fission cross section benefits to all partialcross sections. A sound framework for the simulation of competitive reactions hadto be settled in order to further investigate on the fission reaction; this was implementedusing the TALYS reference code as guideline. After ensuring consistency and consistencyof the framework, focus was made on fission. Perspective resulting from the useof macroscopic-microscopic models such as the FRDM and FRLDM were analyzed; thesemodels have been implemented and validated on experimental data and benchmarks. Tocomply with evaluation requirements in terms of computation time, several specific numericalmethods have been used and parts of the program were written to run on GPU.These macroscopic-microscopic models yield potential energy surfaces that can be used toextract a one-dimensional fission barrier. This latter can then be used to obtained fissiontransmission coefficients that can be used in a Hauser-Feshbach model. This method hasbeen finally tested for the calculation of the average fission cross section for 239Pu(n,f).

<p>Neutron-carbon interactions have been studied with a focus on charged-particle production of relevance to radiation protection and medical applications, such as cancer therapy. The measurements have been performed using the particle-detection setup, MEDLEY, and the 96 MeV neutron beam at the The Svedberg Laboratory in Uppsala.</p><p>Double-differential cross sections of inclusive charged-particle production are compared with recent calculations from models based on the GNASH code including direct, preequilibrium and compound processes. For protons, the shapes of the cross-section spectra are reasonably well described by the calculations. For the other particles- <i>d</i>, <i>t</i>, ³He and α- there are important discrepancies, in particular for ³He-ions and α-particles, concerning both shape and magnitude of the spectra.</p><p>Using the new cross sections, partial as well as total KERMA coefficients have been determined. The coefficients have also been compared to previous experimental results and model calculations. The <i>p</i>, <i>d</i> and <i>t</i> KERMA coefficients are in good agreement with those from a previous measurement. For the helium isotopes, there are no previous measurements at this energy. The KERMA coefficients are considerably higher (by up to 30%) than those predicted by the calculations.</p><p>The KERMA results indicate that protons and α -particles are the main contributors to the dose. A 6x6x6 cm³ carbon phantom, exposed to a broad and a pencil-like beam, is used for the computation of the absorbed doses deposited by these two particles in spheres of 1 μm in diameter, located at various positions in the phantom. The maximum doses are deposited at ~3 cm from the surface of neutron impact for protons and within 1 cm for α-particles. For the pencil beam, deposited doses are spread over regions of ~1.5 cm and ~300 μm transverse to the beam for protons and α-particles, respectively. The results are consistent with previous integral measurements at lower energies.</p>

The collection and interpretation of data is a critical component of traffic and transportation engineering used to establish baseline performance measures and to forecast future conditions. One important source of traffic data is commercial motor vehicle (CMV) weight and classification data used as input to critical tasks in transportation design, operations, and planning. The evolution of Intelligent Transportation System (ITS) technologies has been providing transportation engineers and planners with an increased availability of CMV data. The primary sources of these data are automatic vehicle classification (AVC) and weigh-in-motion (WIM). Microscopic traffic simulation models have been used extensively to model the dynamic and stochastic nature of transportation systems including vehicle composition. One aspect of effective microscopic traffic simulation models that has received increased attention in recent years is the calibration of these models, which has traditionally been concerned with identifying the "best" parameter set from a range of acceptable values. Recent research has begun the process of automating the calibration process in an effort to accurately reflect the components of the transportation system being analyzed. The objective of this research is to develop a methodology in which the effects of CMVs can be included in the calibration of microscopic traffic simulation models. The research examines the ITS data available on weight and operating characteristics of CMVs and incorporates this data in the calibration of microscopic traffic simulation models. The research develops a methodology to model CMVs using microscopic traffic simulation models and then utilizes the output of these models to generate the data necessary to quantify the impacts of CMVs on infrastructure, travel time, and emissions. The research uses advanced statistical tools including principal component analysis (PCA) and recursive partitioning to identify relationships between data collection sites (i.e., WIM, AVC) such that the data collected at WIM sites can be utilized to estimate weight and length distributions at AVC sites. The research also examines methodologies to include the distribution or measures of central tendency and dispersion (i.e., mean, variance) into the calibration process. The approach is applied using the CORSIM model and calibrated utilizing an automated genetic algorithm methodology.

Neutron-carbon interactions have been studied with a focus on charged-particle production of relevance to radiation protection and medical applications, such as cancer therapy. The measurements have been performed using the particle-detection setup, MEDLEY, and the 96 MeV neutron beam at the The Svedberg Laboratory in Uppsala. Double-differential cross sections of inclusive charged-particle production are compared with recent calculations from models based on the GNASH code including direct, preequilibrium and compound processes. For protons, the shapes of the cross-section spectra are reasonably well described by the calculations. For the other particles- d, t, 3He and α- there are important discrepancies, in particular for 3He-ions and α-particles, concerning both shape and magnitude of the spectra. Using the new cross sections, partial as well as total KERMA coefficients have been determined. The coefficients have also been compared to previous experimental results and model calculations. The p, d and t KERMA coefficients are in good agreement with those from a previous measurement. For the helium isotopes, there are no previous measurements at this energy. The KERMA coefficients are considerably higher (by up to 30%) than those predicted by the calculations. The KERMA results indicate that protons and α -particles are the main contributors to the dose. A 6x6x6 cm3 carbon phantom, exposed to a broad and a pencil-like beam, is used for the computation of the absorbed doses deposited by these two particles in spheres of 1 μm in diameter, located at various positions in the phantom. The maximum doses are deposited at ~3 cm from the surface of neutron impact for protons and within 1 cm for α-particles. For the pencil beam, deposited doses are spread over regions of ~1.5 cm and ~300 μm transverse to the beam for protons and α-particles, respectively. The results are consistent with previous integral measurements at lower energies.

Mit der Entwicklung eines stochastischen Modells zur Abbildung des Bewegungsverhaltens von Passagieren wird die Basis für eine virtuelle Anwendungsumgebung geschaffen, mit der die Passagierabfertigungsprozesse im Flughafenterminal und die hierfür notwendigen Infrastrukturen modelliert, implementiert, untersucht und gezielt optimiert werden können. Es werden vorhandene wissenschaftliche Modellansätze zur mikroskopischen Agentensimulation kritisch gewürdigt und Anforderungen an das zu entwickelnde Bewegungsmodell abgeleitet. Das eigens entwickelte stochastische Bewegungsmodell stellt die Erweiterung eines räumlich diskreten mikroskopischen Modells auf Basis eines zellularen Automaten dar, wobei Defizite aufgrund der verwendeten diskreten Gitterstruktur bereits auf Modellebene kompensiert werden. Zu den Erweiterungen zählen die autonome Umgebungsanalyse und die Routenplanung des Agenten, die Abbildung weitreichender Wechselwirkungen zwischen den Agenten und die Berücksichtigung von gruppendynamischen Entscheidungen. Durch die Validierung des stochastischen Bewegungsmodells anhand des Fundamentaldiagramms für Fußgänger kann gezeigt werden, dass das Modell in der Lage ist, den charakteristischen Verlauf der Geschwindigkeit in Relation zur Agentendichte quantitativ abzubilden. Auch typische, in der Realität beobachtbare Selbstorganisationseffekte können durch das Modell reproduziert werden. Für die Anwendung des stochastischen Modells zur Abbildung des Bewegungsverhaltens von Passagieren im Flughafenterminal wird das Modell durch empirisch erhobene Passagierbewegungsdaten kalibriert. Die Datenerhebung erfolgt am Flughafen Dresden unter Verwendung eines entwickelten videogestützten Bewegungsverfolgungssystems und erlaubt eine gezielte Kalibrierung hinsichtlich der Passagierparameter: Geschlecht, Reisemotivation (privat oder geschäftlich), Gruppengröße sowie Gepäckart und -anzahl. Für die Erstellung der virtuellen Terminalumgebung werden die Passagierabfertigungsprozesse eingehend analysiert und die Prozesszeiten der jeweiligen Abfertigungsstationen durch spezifische Wahrscheinlichkeitsverteilungen modelliert. Hierfür stehen empirische Datenerhebungen am Flughafen Stuttgart zur Verfügung, die eine detaillierte Prozessanalyse hinsichtlich der Passagierparameter und der Prozessparameter (Erfahrung des Personals, Reaktionszeiten bei Störungen) erlauben. Im Anschluss an die Kalibrierung des stochastischen Bewegungsmodells und die Modellierung der Passagierabfertigungsprozesse erfolgt die Entwicklung einer Anwendungsumgebung für die Implementierung des virtuellen Flughafens. Durch den modularen Aufbau der Anwendungsumgebung ist eine effiziente Implementierung der Flughafenstrukturen (Grundriss, Flugplan, Personaleinsatz), der Abfertigungsprozesse und des stochastischen Bewegungsmodells möglich. Die Anwendungsumgebung stellt dabei einen übergeordneten Rahmen dar, durch den eine allgemeine Nutzerschnittstelle (Konfigurationsumgebung), eine grafische Ergebnisaufbereitung und die dreidimensionale Abbildung des Bewegungsverhaltens der Passagiere zur Verfügung steht. Die Anwendung des entwickelten stochastischen Bewegungsmodells erfolgt für die Validierung der Passagierabfertigungsprozesse (Check-In und Sicherheitskontrolle), für die Entwicklung einer passagierbezogener Prozessbewertung und für die vollständige Abbildung der Terminalprozesse (Abflug) am Beispiel des Flughafens Dresden. Durch die Analyse des Einstiegsverhaltens der Passagiere in ein Verkehrsflugzeug werden die Notwendigkeit des Einsatzes stochastischer Bewegungsmodelle und das Potential mikroskopischer Modellierungsansätze verdeutlicht. Das entwickelte stochastische Bewegungsmodell kann das Passagierverhalten auch in komplexen Umgebungen umfänglich widerspiegeln und die entwickelte Anwendungsumgebung stellt einen idealen Rahmen für die Modellanwendung und -weiterentwicklung dar. Durch die anwendungsorientierten Implementierungen steht eine Vielzahl von geeigneten Detaillösungen zur Verfügung, um den zukünftigen wissenschaftlichen und praxisrelevanten Herausforderungen der Personendynamik zu begegnen. / The development of a stochastic motion model allows for using a virtual application environment, to reproduce passenger motion behavior and handling processes at airport terminals. Based on the introduced scientific approaches for microscopic agent simulation, requirements for an application-oriented motion model are derived. The developed model is a substantial extension of a stochastic cellular automata approach, where the deficiencies due to the discrete grid structure are compensated on a fundamental level. The model development is completed by adding agent-oriented environment analysis, route planning, and mid-range agent interaction. The stochastic motion model proves its capabilities for a quantitative reproduction of the characteristic shape of the common fundamental diagram of pedestrian dynamics. Moreover, generic self-organization effects are reproduced by the model. For the application of the stochastic approach for modeling the motion behavior of passengers inside an airport terminal, a comprehensive acqusition of data at Dresden International Airport provides a solid basis. A video-supported tracking environment allows for an efficient categorization of passengers and analysis of their motion behavior regarding to their gender, travel purpose (private or business), group size, and baggage types and quantities. In addition to the passenger-related data, the process time of passenger handling at each station at Stuttgart Airport is analyzed in detail and transformed to statistic probabilities by functional data fitting. Finally, the calibrated stochastic motion model is prepared for passenger dynamics at airport terminals. After the successful development and calibration, the implementation of the motion model in a virtual application environment is accomplished. To implement the terminal structure, the passenger handling processes, and the individual passenger motion behavior common programming interfaces are used as well as specific components for linking model and animation requirements. The application of the stochastic motion models aims at the validation of passenger handling process on the basis of empirical data from Stuttgart airport and at the development of a passenger-oriented process evaluation using Dresden Airport environment. The simulation of passenger dynamics at airport terminals points out that the stochastic motion model reproduces the motion behavior of passengers close to reality. Due to the application-oriented implementation a variety of appropriate solutions are available for future scientific and operational challenges related to passenger dynamics.

The collection and interpretation of data is a critical component of traffic and transportation engineering used to establish baseline performance measures and to forecast future conditions. One important source of traffic data is commercial motor vehicle (CMV) weight and classification data used as input to critical tasks in transportation design, operations, and planning. The evolution of Intelligent Transportation System (ITS) technologies has been providing transportation engineers and planners with an increased availability of CMV data. The primary sources of these data are automatic vehicle classification (AVC) and weigh-in-motion (WIM). Microscopic traffic simulation models have been used extensively to model the dynamic and stochastic nature of transportation systems including vehicle composition. One aspect of effective microscopic traffic simulation models that has received increased attention in recent years is the calibration of these models, which has traditionally been concerned with identifying the "best" parameter set from a range of acceptable values. Recent research has begun the process of automating the calibration process in an effort to accurately reflect the components of the transportation system being analyzed. The objective of this research is to develop a methodology in which the effects of CMVs can be included in the calibration of microscopic traffic simulation models. The research examines the ITS data available on weight and operating characteristics of CMVs and incorporates this data in the calibration of microscopic traffic simulation models. The research develops a methodology to model CMVs using microscopic traffic simulation models and then utilizes the output of these models to generate the data necessary to quantify the impacts of CMVs on infrastructure, travel time, and emissions. The research uses advanced statistical tools including principal component analysis (PCA) and recursive partitioning to identify relationships between data collection sites (i.e., WIM, AVC) such that the data collected at WIM sites can be utilized to estimate weight and length distributions at AVC sites. The research also examines methodologies to include the distribution or measures of central tendency and dispersion (i.e., mean, variance) into the calibration process. The approach is applied using the CORSIM model and calibrated utilizing an automated genetic algorithm methodology.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais / Leptospirosis is a zoonosis of global public health importance and can affect all groups of vertebrates, including reptiles. These animals can play a role in the epidemiological cycle of the disease spreading and keeping the causative agent in the environment. Leptospirosis is little known in snakes and this study was designed to cover the gaps in the occurrence of the disease in these animals. The occurrence of anti-Leptospira spp. antibodies in rattlesnakes Crotalus durissus collilineatus kept in captivity was evaluated and the most common serotypes in these animals were determined using the Microscopic Agglutination Test (MAT). Changes in concentrations of plasma biochemical constituents in positive snakes were also evaluated, depending on the antibody titre presented at MAT. Almost 90% of the rattlesnakes used in this study were positive to the serological test and the most common serotypes were Javanica, Andamana and Patoc. The most common titres were 25 and 50, but there were titres as high as 1600. Captive snakes can act as important sources of leptospira infection for humans and so it is critical that the disease is recognized and controlled. Although apparently healthy, alterations in biochemical parameters indicated inflammation and infection in animals with low antibody titres. Rattlesnakes with high titres showed signs of severe kidney damage and liver disease, which are common in leptospirosis. The biochemical assessment of positive MAT reptiles may be important to diagnose the infection in these animals. / A leptospirose é uma zoonose de importância em saúde pública global e pode acometer praticamente todos os grupos de vertebrados, inclusive répteis. Esses animais podem atuar no ciclo epidemiológico da doença mantendo e disseminando o agente causador no ambiente. A leptospirose é pouco conhecida em serpentes e, por isso, objetivou-se estudar a ocorrência da doença nesses animais. Para isso, foi avaliada a ocorrência de anticorpos anti-Leptospira spp. em Crotalus durissus collilineatus mantidas em cativeiro e as sorovariedades mais frequentes nesses animais, utilizando o teste de Soroaglutinação Microscópica (SAM). Também foram avaliadas as alterações nas concentrações de constituintes bioquímicos plasmáticos nas serpentes reagentes, relacionadas com o título de anticorpos apresentado na SAM. Quase 90% das cascavéis utilizadas nesse estudo foram reagentes no teste sorológico com triagem de 1:25 e os sorovares mais frequentes foram Javanica, Andamana e Patoc. Os títulos mais frequentes foram 25 e 50, mas houve títulos de até 1600. Apesar de aparentemente saudáveis, a avaliação dos parâmetros bioquímicos indicou inflamação e infecção mesmo nos animais com títulos de anticorpos considerados baixos para mamíferos (&#8804;50). As cascavéis com títulos altos apresentaram quadros mais graves que aquelas com títulos menores, sugestivos de lesão renal grave e indícios de doença hepática, que são lesões características da leptospirose. Serpentes de cativeiro podem atuar como fontes de infecção de leptospiras para humanos e outros animais e por isso, é fundamental que a doença seja conhecida e controlada em criadouros. A avaliação bioquímica de répteis reagentes no teste sorológico pode ser importante para o diagnóstico da infecção por Leptospira spp. nesses animais. / Mestre em Ciências Veterinárias