Dissertationen zum Thema „Source term estimation“

Conditional Source-term Estimation (CSE) methods are used to obtain chemical closure in turbulent combustion simulation. A Laminar Flamelet Decomposition (LFD) and then a Trajectory Generated Low-Dimensional Manifold (TGLDM) method are combined with CSE in Reynolds-Averaged Navier Stokes (RANS) simulation of non-premixed autoigniting jets. Despite the scatter observed in the experimental data, the predictions of ignition delay from both methods agree reasonably well with the measurements. The discrepancy between predictions of these two methods can be attributed to different ways of generating libraries that contain information of detailed chemical mechanism. The CSE-TGLDM method is recommended for its seemingly better performance and its ability to transition from autoignition to combustion. The effects of fuel composition and injection parameters on ignition delay are studied using the CSE-TGLDM method. The CSE-TGLDM method is then applied in Large Eddy Simulation of a non-premixed, piloted jet flame, Sandia Flame D. The adiabatic CSE-TGLDM method is extended to include radiation by introducing a variable enthalpy defect to parameterize TGLDM manifolds. The results are compared to the adiabatic computation and the experimental data. The prediction of NO formation is improved, though the predictions of temperature and major products show no significant difference from the adiabatic computation due to the weak radiation of the flame. The scalar fields are then extracted and used to predict the mean spectral radiation intensities of the flame. Finally, the application of CSE in turbulent premixed combustion is explored. A product-based progress variable is chosen for conditioning. Presumed Probability Density Function (PDF) models for the progress variable are studied. A modified version of a laminar flame-based PDF model is proposed, which best captures the distribution of the conditional variable among all PDFs under study. A priori tests are performed with the CSE and presumed PDF models. Reaction rates of turbulent premixed flames are closed and compared to the DNS data. The results are promising, suggesting that chemical closure can be achieved in premixed combustion using the CSE method.

Conditional Source-term Estimation (CSE) is a closure model for turbulence-chemistry interactions. This model is based on the conditional moment closure hypothesis for the chemical reaction source terms. The conditional scalar field is estimated by solving an integral equation using inverse methods. CSE was originally developed for - and has been used extensively in - non-premixed combustion. This work is the first application of this combustion model to predictive simulations of turbulent premixed flames. The underlying inverse problem is diagnosed with rigorous mathematical tools. CSE is coupled with a Trajectory Generated Low-Dimensional Manifold (TGLDM) model for chemistry. The CSE-TGLDM combustion model is used with both Reynolds-Averaged Navier-Stokes (RANS) and Large-Eddy Simulation (LES) turbulence models to simulate two different turbulent premixed flames. Also in this work, the Presumed Conditional Moment (PCM) turbulent combustion model is employed. This is a simple flamelet model which is used with the Flame Prolongation of ILDM (FPI) chemistry reduction technique. The PCM-FPI approach requires a presumption for the shape of the probability density function of reaction progress variable. Two shapes have been examined: the widely used beta-function and the Modified Laminar Flamelet PDF (MLF-PDF). This model is used in both RANS and large-eddy simulation of a turbulent premixed Bunsen burner. Radial distributions of the calculated temperature field, axial velocity and chemical species mass fraction have been compared with experimental data. This comparison shows that using the MLF-PDF leads to predictions that are similar, and often superior to those obtained using the beta-PDF. Given that the new PDF is based on the actual chemistry - as opposed to the ad hoc nature of the beta-PDF - these results suggest that it is a better choice for the statistical description of the reaction progress variable.

Conditional Source-term Estimation (CSE) is a chemical closure model for the simulation of turbulent combustion. In this work, CSE has been explored for modelling combustion phenomena in a spark-ignition (SI) engine. In the arbitrarily complex geometries imposed by industrial design, estimation of conditionally averaged scalars is challenging. The key underlying requirement of CSE is that conditionally averaged scalars be calculated within spatially localized sub-domains. A domain partitioning algorithm based on space-filling curves has been developed to construct localized ensembles of points necessary to retain the validity of CSE. Algorithms have been developed to evenly distribute points to the maximum extent possible while maintaining spatial locality. A metric has been defined to estimate relative inter-partition contact as an indicator of communication in parallel computing architectures. Domain partitioning tests conducted on relevant geometries highlight the performance of the method as an unsupervised and computationally inexpensive domain partitioning tool. In addition to involving complex geometries, SI engines pose the challenge of accurately modelling the transient ignition process. Combustion in a homogeneous-charge natural gas fuelled SI engine with a relatively simple chamber geometry has been simulated using an empirical model for ignition. An oxygen based reaction progress variable is employed as the conditioning variable and its stochastic behaviour is approximated by a presumed probability density function (PDF). A trajectory generated low-dimensional manifold has been used to tabulate chemistry in a hyper-dimensional space described by the reaction progress variable, temperature and pressure. The estimates of pressure trace and pollutant emission trends obtained using CSE accurately match experimental measurements.

Combustion modeling is performed with Conditional Source-term Estimation (CSE) using both Laminar Flamelet Decomposition (LFD) and Low-dimensional Manifolds. CSE with Laminar Flamelet Decomposition (LFD) is used in the Large Eddy Simulation (LES) context to study the non-premixed Sandia D-flame. The results show that the flame temperature and major species are well predicted with both steady and unsteady flamelet libraries. A mixed library composed of steady and unsteady flamelet solutions is needed to get a good prediction of NO. That the LFD model allows for tuning of the results is found to be significant drawback to this approach. CSE is also used with a Trajectory Generated Low-dimensional Manifold (TGLDM) to simulate the Sandia D-flame. Both GRI-Mech 3.0 and GRI-Mech 2.11 are found to be able to predict the temperature and major species well. However, only GRI-Mech 2.11 gives a good prediction of NO. That GRI-Mech 3.0 failed to give a good prediction of NO is in agreement with the findings of others in the literature. The Stochastic Particle Model (SPM) is used to extend the TGLDM to low temperatures where the original continuum TGLDM failed. A new method for generating a trajectory for the TGLDM by averaging different realizations together is proposed. The new TGLDM is used in simulations of a premixed laminar flame and a perfectly stirred reactor. The results show that the new TGLDM significantly improves the prediction. Finally, a time filter is applied to individual SPM realizations to eliminate the small time scales. These filtered realizations are tabulated into TGLDM which are then used to predict the autoignition delay time of a turbulent methane/air jet in RANS using CSE. The results are compared with shock tube experimental data. The TGLDMs incorporating SPM results are able to predict a certain degree of fluctuations in the autoignition delay time, but the magnitude is smaller than is seen in the experiments. This suggests that fluctuations in the ignition delay are at least in part due to turbulent fluctuations, which might be better predicted with LES.
Science, Faculty of
Mathematics, Department of
Graduate

This work evaluates whether the branch of Reynolds Averaging in Computational Fluid Dynamics can be used to, based on real field measurements, find the source of the measured gas in question. The method to do this is via the adjoint to the Reynolds Averaged Scalar Transport equation, explained and derived herein. Since the Inverse is only as good as the main equation, forward runs are made to evaluate the turbulence model. Reynolds Averaged Navier Stokes is solved in a domain containing 4 cubes in a 2x2 grid, generating a velocity field for said domain. The turbulence model in question is a union of two modifications to the standard two equation k-ε model in order to capture blunt body turbulence but also to model the atmospheric boundary layer. This field is then inserted into the Reynolds Averaged Scalar Transport equation and the simulation is compared to data from the Environmental Flow wind tunnel in Surrey. Finally the adjoint scalar transport is solved, both for synthetic data that was generated in the forward run, but also for the data from EnFlo. It was discovered that the turbulent Schmidt number plays a major role in capturing the dispersed gas, three different Schmidt numbers were tested, the standard 0.7, the unconventional 0.3 and a height dependent Schmidt number. The widely accepted value of 0.7 did not capture the dispersion at all and gave a huge model error. As such the adjoint scalar transport was solved for 0.3 and a height dependent Schmidt number. The interaction between measurements, the real source strength (which is not used in the adjoint equation, but needed to find the source) and the location of the source is intricate indeed. Over estimation and under estimation of the forward model may cancel out in order to find the correct source, with the correct strength. It is found that Reynolds Averaged Computational fluid dynamics may prove useful in source term estimation.
Detta arbete utvärderar hurvida Reynolds medelvärdesmodellering inom flödessimuleringar kan användas till att finna källan till en viss gas baserat på verkliga mätningar ute i fält. Metoden går ut på att använda den adjungerade ekvationen till Reynolds tidsmedlade skalära transportekvationen, beskriven och härledd häri. Då bakåtmodellen bygger på framåtmodellen, måste såleds framåtmodellen utvärderas först. Navier-Stokes ekvationer med en turbulensmodell löses i en domän, innehållandes 4 kuber i en 2x2 orientering, för vilken en hastighetsprofil erhålles. Turbulensmodellen som användes är en union av två olika k-ε modeller, där den ena fångar turbulens runt tröga objekt och den andra som modellerar atmosfäriska gränsskiktet. Detta fält används sedan i framåtmodellen av skalära transportekvationen, som sedan jämförs med körningar från EnFlo windtunneln i Surrey. Slutligen testkörs även den adjungerade ekvationen, både för syntetiskt data genererat i framåtkörningen men även för data från EnFlo tunneln. Då det visade sig att det turbulenta Schmidttalet spelar stor roll inom spridning i det atmosfäriska gränsskiktet, gjordes testkörningar med tre olika Schmidttal, det normala 0.7, det väldigt låga talet 0.3 samt ett höjdberoende Schmidttal. Det visade sig att det vanligtvis använda talet 0.7 inte alls lyckas fånga spridningen tillfredställande och gav ett stort modellfel. Därför löstes den adjungerade ekvationen för 0.3 samt för ett höjdberoende Schmidttal. Interaktionen mellan mätningar, den riktiga källstyrkan (som är okänd i den adjungerade ekvationen) samt källpositionen är onekligen intrikat. Över- samt underestimationer av framåtmodellen kan ta ut varandra i bakåtmodellen för att finna rätt källa, med rätt källstyrka. Det ter sig som Reynolds turbulensmodellering mycket möjligt kan användas inom källtermsuppskattning.

Computational fluid dynamics (CFD) is indispensable in the development of complex engines due to its low cost and time requirement compared to experiments. Nevertheless, because of the strong coupling between turbulence and chemistry in premixed flames, the prediction of chemical reaction source terms continues to be a modelling challenge. This work focuses on the improvement of turbulent premixed combustion simulation strategies requiring the use of presumed probability density function (PDF) models. The study begins with the development of a new PDF model that includes the effect of turbulence, achieved by the implementation of the Linear-Eddy Model (LEM). Comparison with experimental burners reveals that the LEM PDF can capture the general PDF shapes for methane-air combustion under atmospheric conditions with greater accuracy than other presumed PDF models. The LEM is additionally used to formulate a new, pseudo-turbulent scalar dissipation rate (SDR) model. Conditional Source-term Estimation (CSE) is implemented in the Large Eddy Simulation (LES) of the Gülder burner as the closure model for the chemistry-turbulence interactions. To accommodate the increasingly parallel computational environments in clusters, the CSE combustion module has been parallelised and optimised. The CSE ensembles can now dynamically adapt to the changing flame distributions by shifting their spatial boundaries and are no longer confined to pre-allocated regions in the simulation domain. Further, the inversion calculation is now computed in parallel using a modified version of an established iterative solver, the Least-Square QR-factorisation (LSQR). The revised version of CSE demonstrates a significant reduction in computational requirement — a reduction of approximately 50% — while producing similar solutions as previous implementations. The LEM formulated PDF and SDR models are subsequently implemented in conjunction with the optimised version of CSE for the LES of a premixed methane-air flame operating in the thin reaction zone. Comparison with experimental measurements of temperature reveals that the LES results are very comparable in terms of the flame height and distribution. This outcome is encouraging as it appears that this work represents a significant step towards the correct direction in developing a complete combustion simulation strategy that can accurately predict flame characteristics in the absence of ad hoc parameters.
Applied Science, Faculty of
Mechanical Engineering, Department of
Graduate

Depuis l'ère industrielle, les villes sont impactées par la pollution de l'air du fait de la densité de l'industrie, de la circulation de véhicules et de la densité d'appareils de chauffage à combustion. La pollution atmosphérique urbaine a des conséquences sur la santé auxquelles les pouvoirs publics et les citoyens s'intéressent toujours plus. Cette pollution peut entraîner l'aggravation de l'asthme ou encore de troubles cardiovasculaires.Le but de ce travail de thèse est de localiser et de quantifier des sources de pollution urbaine à l'aide d'un réseau dense de mesures bruitées. Nous avons fait le choix de développer des méthodes d'estimation de sources de pollution en se basant sur des modèles physiques de dispersion de polluants. Ainsi l'estimation de sources de pollution est contrainte par la connaissance de la physique du phénomène de dispersion.Ce travail de thèse porte ainsi sur la modélisation numérique de la dispersion de polluants en milieu urbain et sur l'estimation des termes sources.À cause des nombreuses contraintes qu'impose le bâti urbain aux flux de polluants, la physique de dispersion est représentée par des modèles numériques coûteux en calcul.Nous avons développé un modèle numérique de dispersion basé sur la méthode des Différences Finies supportées par des Fonctions de Bases Radiales (RBF-FD). Ces approches sont réputées frugales en calcul et adaptées au traitement de domaines de simulation à géométrie complexe. Notre modèle RBF-FD peut traiter des problèmes bidimensionnels (2D) et tridimensionnels (3D). Nous avons comparé cette modélisation face à un modèle analytique en 2D, et avons qualitativement comparé notre modèle en 3D à un modèle numérique de référence.Des expériences d'estimations de source ont ensuite été réalisées. Elles considèrent de nombreuses mesures bruitées pour estimer un terme source quelconque sur tout le domaine de simulation. Les différentes études effectuées mettent en oe uvre des expériences jumelles : nous générons nous-même des mesures simulées par un modèle numérique et évaluons les performances des estimations. Après avoir testé une approche d'apprentissage automatique sur un cas unidimensionnel en régime stationnaire, nous avons testé des méthodes d'estimation de terme source sur des cas tridimensionnels en régime permanent et transitoire, en considérant des géométries sans et avec présence d'obstacles. Nous avons testé des estimations en utilisant une méthode adjointe originale puis une méthode originale d'estimation inspirée de l'apprentissage automatique informé de la physique (PIML) et enfin un filtre de Kalman. L'approche inspirée du PIML pour le moment testée en régime stationnaire conduit à une qualité d'estimation comparable au filtre de Kalman (où ce dernier considère un régime transitoire de dispersion à source stationnaire). L'approche inspirée du PIML exploite directement la frugalité du modèle direct de calcul RBF-FD, ce qui en fait une méthode prometteuse pour des estimations de source sur des domaines de calcul de grande taille
Since the industrial era, cities have been affected by air pollution due to the density of industry, vehicle traffic and the density of combustion heaters. Urban air pollution has health consequences that are of increasing concern to both public authorities and the general public. This pollution can aggravate asthma and cardiovascular problems. The aim of this thesis is to locate and quantify sources of urban pollution using a dense network of noisy measurements. We have chosen to develop methods for estimating pollution sources based on physical models of pollutant dispersion. The estimation of pollution sources is therefore constrained by knowledge of the physics of the dispersion phenomenon. This thesis therefore focuses on the numerical modelling of pollutant dispersion in an urban environment and on the estimation of source terms.Because of the many constraints imposed on pollutant flows by urban buildings, the physics of dispersion is represented by computationally expensive numerical models.We have developed a numerical dispersion model based on the Finite Difference method supported by Radial Basis Functions (RBF-FD). These approaches are known to be computationally frugal and suitable for handling simulation domains with complex geometries. Our RBF-FD model can handle both two-dimensional (2D) and three-dimensional (3D) problems. We compared this model with a 2D analytical model, and qualitatively compared our 3D model with a reference numerical model.Source estimation experiments were then carried out. They consider numerous noisy measurements in order to estimate any source term over the entire simulation domain. The various studies carried out involve twin experiments: we ourselves generate measurements simulated by a numerical model and evaluate the performance of the estimates. After testing a machine-learning approach on a one-dimensional steady-state case, we tested source term estimation methods on three-dimensional steady-state and transient cases, considering geometries without and with the presence of obstacles. We tested estimates using an original adjoint method, then an original estimation method inspired by physics-informed machine learning (PIML) and finally a Kalman filter. The PIML-inspired approach, which is currently being tested in a stationary regime, produces an estimation quality comparable to that of the Kalman filter (where the latter considers a transient dispersion regime with a stationary source). The PIML-inspired approach directly exploits the frugality of the RBF-FD direct computation model, which makes it a promising method for source estimates over large computational domains

En physique de l’atmosphère, la reconstruction d’une source polluante à partir des mesures de capteurs est une question importante. Elle permet en effet d’affiner les paramètres des modèles de dispersion servant à prévoir la propagation d’un panache de polluant, et donne aussi des informations aux primo-intervenants chargés d’assurer la sécurité des populations. Plusieurs méthodes existent pour estimer les paramètres de la source, mais leur application est coûteuse à cause de la complexité des modèles de dispersion. Toutefois, cette complexité est souvent nécessaire, surtout lorsqu’il s’agit de traiter des cas urbains où la présence d’obstacles et la météorologie instationnaire imposent un niveau de précision important. Il est aussi vital de tenir compte des différents facteurs d’incertitude, sur les observations et les estimations. Les travaux menés dans le cadre de cette thèse ont pour objectif de développer une méthodologie basée sur l’inférence bayésienne adaptative couplée aux méthodes de Monte Carlo pour résoudre le problème d’estimation du terme source. Pour cela, nous exposons d’abord le contexte scientifique du problème et établissons un état de l’art. Nous détaillons ensuite les formulations utilisées dans le cadre bayésien, plus particulièrement pour les algorithmes d’échantillonnage d’importance adaptatifs. Le troisième chapitre présente une application de l’algorithme AMIS dans un cadre expérimental, afin d’exposer la chaîne de calcul utilisée pour l’estimation de la source. Enfin, le quatrième chapitre se concentre sur une amélioration du traitement des calculs de dispersion, entraînant un gain important de temps de calcul à la fois en milieu rural et urbain
In atmospheric physics, reconstructing a pollution source is a challenging but important question : it provides better input parameters to dispersion models, and gives useful information to first-responder teams in case of an accidental toxic release.Various methods already exist, but using them requires an important amount of computational resources, especially as the accuracy of the dispersion model increases. A minimal degree of precision for these models remains necessary, particularly in urban scenarios where the presence of obstacles and the unstationary meteorology have to be taken into account. One has also to account for all factors of uncertainty, from the observations and for the estimation. The topic of this thesis is the construction of a source term estimation method based on adaptive Bayesian inference and Monte Carlo methods. First, we describe the context of the problem and the existing methods. Next, we go into more details on the Bayesian formulation, focusing on adaptive importance sampling methods, especially on the AMIS algorithm. The third chapter presents an application of the AMIS to an experimental case study, and illustrates the mechanisms behind the estimation process that provides the source parameters’ posterior density. Finally, the fourth chapter underlines an improvement of how the dispersion computations can be processed, thus allowing a considerable gain in computation time, and giving room for using a more complex dispersion model on both rural and urban use cases

Nous analysons dans cette thèse différents aspects associés à la modélisation des écoulements à surface libre en eaux peu profondes (Shallow Water). Nous étudions tout d’abord le système d’équations de Saint-Venant à deux dimensions et leur résolution par la méthode numérique des volumes finis, en portant une attention particulière sur les aspects hyperboliques et conservatifs. Ces schémas permettent de traiter les équilibres stationnaires, les interfaces sec/mouillé et aussi de modéliser des écoulements subcritique, transcritique et supercritique. Nous présentons ensuite la théorie de la méthode d’assimilation variationnelle de données adaptée à ce type d’écoulement. Son application au travers des études de sensibilité est longuement discutée dans le cadre de l'hydraulique à surface libre. Après cette partie à caractère théorique, la partie tests commence par une qualification de l’ensemble des méthodes numériques qui sont implémentées dans le code DassFlow, développé à l’Université de Toulouse, principalement à l’IMT mais aussi à l’IMFT. Ce code résout les équations Shallow Water par une méthode de volumes finis et est validé par comparaison avec les solutions analytiques pour des cas tests classiques. Ces mêmes résultats sont comparés avec un autre code d’hydraulique à surface libre aux éléments finis en deux dimensions, Telemac 2D. Une particularité notable du code DassFlow est de permettre l’assimilation variationnelle de données grâce au code adjoint permettant le calcul du gradient de la fonction coût. Ce code adjoint a été obtenu en utilisant l'outil de différentiation automatique Tapenade (Inria). Nous testons ensuite sur un cas réel, hydrauliquement complexe, différentes qualités de Modèles Numériques de Terrain (MNT) et de bathymétrie du lit d’une rivière. Ces informations proviennent soit d’une base de données classique type IGN, soit d’informations LIDAR à très haute résolution. La comparaison des influences respectives de la bathymétrie, du maillage et du type de code utilisé, sur la dynamique d’inondation est menée très finement. Enfin nous réalisons des études cartographiques de sensibilité aux paramètres du modèle sur DassFlow. Ces cartes montrent l’influence respective des différents paramètres ou de la localisation des points de mesure virtuels. Cette localisation optimale de ces points est nécessaire pour une future assimilation de données efficiente
We analyze in this thesis various aspects associated with the modeling of free surface flows in shallow water approximation. We first study the system of Saint-Venant equations in two dimensions and its resolution with the numerical finite volumes method, focusing in particular on aspects hyperbolic and conservative. These schemes can process stationary equilibria, wetdry interfaces and model subcritical, transcritical and supercritical flows. After, we present the variational data assimilation method theory fitted to this kind of flow. Its application through sensitivity studies is fully discussed in the context of free surface water. After this theoretical part, we test the qualification of numerical methods implemented in the code Dassflow, developed at the University of Toulouse, mainly at l'IMT, but also at IMFT. This code solves the Shallow Water equations by finite volume method and is validated by comparison with analytical solutions for standard test cases. These results are compared with another hydraulic free surface flow code using finite elements in two dimensions: Telemac2D. A significant feature of the Dassflow code is to allow variational data assimilation using the adjoint method for calculating the cost function gradient. The adjoint code was obtained using the automatic differentiation tool Tapenade (INRIA). Then, the test is carried on a real hydraulically complex case using different qualities of Digital Elevation Models (DEM) and bathymetry of the river bed. This information are provided by either a conventional database types IGN or a very high resolution LIDAR information. The comparison of the respective influences of bathymetry, mesh size, kind of code used on the dynamics of flooding is very finely explored. Finally we perform sensitivity mapping studies on parameters of the Dassflow model. These maps show the respective influence of different parameters and of the location of virtual measurement points. This optimal location of these points is necessary for an efficient data assimilation in the future

Cette thèse porte sur l'estimation simultanée du coefficient de diffusion et du terme source régissant le modèle de transport de la température dans les plasmas chauds. Ce phénomène physique est décrit par une équation différentielle partielle (EDP) linéaire, parabolique du second-ordre et non-homogène, où le coefficient de diffusion est distribué et le coefficient de réaction est constant. Ce travail peut se présenter en deux parties. Dans la première, le problème d'estimation est traité en dimension finie ("Early lumping approach"). Dans la deuxième partie, le problème d'estimation est traité dans le cadre initial de la dimension infinie ("Late lumping approach"). Pour l'estimation en dimension finie, une fois le modèle établi, la formulation de Galerkin et la méthode d'approximation par projection sont choisies pour convertir l'EDP de transport en un système d'état linéaire, temps-variant et à entrées inconnues. Sur le modèle réduit, deux techniques dédiées à l'estimation des entrées inconnues sont choisies pour résoudre le problème. En dimension infinie, l'estimation en-ligne adaptative est adoptée pour apporter des éléments de réponse aux contraintes et limitations dues à la réduction du modèle. Des résultats de simulations sur des données réelles et simulées sont présentées dans ce mémoire.

Submitted by Pedro Silva Filho (pfsilva@ipen.br) on 2017-11-21T11:45:58Z No. of bitstreams: 0
Made available in DSpace on 2017-11-21T11:45:58Z (GMT). No. of bitstreams: 0
O gerenciamento de riscos em uma instalação nuclear é necessário para a segurança de trabalhadores e de populações vizinhas. Parte desse processo é a comunicação dos riscos que propicia o diálogo entre gestores da empresa e moradores das áreas de risco. A população que conhece os riscos a que está exposta, como esses riscos são gerenciados e o que deve ser feito em uma situação de emergência tende a se sentir mais segura e a confiar nas instituições responsáveis pelo plano de emergência. Sem diálogo entre empresa e público, o conhecimento dos procedimentos a serem seguidos em caso de acidente não chega à população, ou quando chega, não há confiança dessas pessoas na sua eficácia. Em Angra dos Reis, no litoral sul do Estado do Rio de Janeiro, está a Central Nuclear Almirante Álvaro Alberto. No entorno dessa Central Nuclear existe uma população que, de acordo com o Plano de Emergência Externo (PEE/RJ), deverá ser evacuada ou ficar abrigada, caso ocorra um acidente na instalação. Um trabalho de comunicação de riscos entre esses moradores é necessário para que eles conheçam o plano de emergência e os procedimentos corretos para uma situação de emergência, além de buscar esclarecer dúvidas e mitos. Esse trabalho apresenta uma análise da comunicação dos riscos feita para a população local, a percepção que ela tem dos riscos e o grau de conhecimento do plano de emergência externo por parte dessas pessoas.
Tese (Doutorado em Tecnologia Nuclear)
IPEN/T
Instituto de Pesquisas Energéticas e Nucleares - IPEN-CNEN/SP

L’impact anthropique sur les cours d’eau a significativement augmenté suite à la révolution industrielle engagée par les pays occidentaux. Ainsi, les modifications de la géomorphologie des cours d’eau pour le stockage de l’eau et la navigation, la conversion des surfaces à des fins agricoles, industrielles et d’urbanisation illustrent cette pression environnementale, qui se traduit, en autre, par une augmentation de rejets de divers contaminants dans les compartiments environnementaux et notamment les rivières. Une part de ces rejets peut donc se retrouver dans les matières en suspension, considérées alors comme des puits de stockage, qui transitent dans les rivières. Les aménagements des rivières et notamment la construction de barrages favorisent alors la sédimentation de ces particules contaminées au cours du temps. Ces sédiments d’origines anthropiques, également appelés legacy sediments, sont donc les témoins des activités anthropiques et permettent de reconstruire les trajectoires temporelles des contaminations au sein des bassins versants. L’Eure, affluent majeur de l’estuaire de Seine, a connu d’importantes pressions anthropiques depuis le vingtième siècle. La reconstruction temporelle des pressions anthropiques a nécessité l’association de différentes approches méthodologiques : (i) une analyse diachronique des modifications morphologiques de la rivière a été menée, conjointement à (ii) une analyse de la dynamique sédimentaire et de la nature des dépôts sédimentaires par couplage de méthodes géophysiques, sédimentologiques et géochimiques, et à (iii) la mise en place d’un réseau de suivi du comportement hydro-sédimentaire avec un échantillonnage en continu des matières en suspensions. De profondes modifications géomorphologiques se sont produites en aval du bassin versant, avec pour principales conséquences un exutoire déplacé d’une dizaine de kilomètres en direction d’un barrage et la formation d’annexes hydrauliques favorisant l’accumulation de sédiments dès les années 1940. Ceux-ci ont permis de montrer que le bassin versant de l’Eure avait connu d’importantes contaminations dont les conséquences sont encore enregistrées malgré l’arrêt des activités ou usages. Les tendances temporelles des éléments traces métalliques et métalloïdes ont montré de fortes contaminations en As dans les années 1940 et des contaminations d’origines industrielles en Cr, Co, Ni, Cu, Zn, Ag et Cd durant les années 1960–1970, ainsi que des contaminations en Sb et Pb en 1990–2000. Ces dernières sont toujours enregistrées malgré l’arrêt des activités responsables des rejets, comme l’ont attesté les résultats issus des matières en suspension actuellement collectées dans le cours d’eau. A l’instar d’une majorité des métaux traces, les contaminants organiques, tels les HAPs, ont montré d’importantes contaminations durant les années 1940–1960, dont les signatures indiquent une origine majoritairement pyrogénique. Les PCBs ont montré des contaminations importantes lors de la période 1950–1970, en lien avec la production et les usages nationaux de mélanges composés en majorité de congénères faiblement chlorés. Enfin, l’intérêt porté à une troisième famille de contaminants organiques persistants, les pesticides organochlorés, a montré l’utilisation de lindane et du DDT notamment lors de la période 1940–1970, et a mis en avant d’une part une utilisation post-interdiction du lindane et d’autre part la présence d’un métabolite du DDT plusieurs décennies après l’arrêt d’utilisation de ce dernier, en lien avec l’augmentation de l’érosion des sols cultivés
The anthropogenic impact on rivers has significantly increased following the industrial revolutioninitiated by Western countries. Thus, changes in the geomorphology of rivers for water storage andnavigation, the conversion of land for agricultural, industrial and urbanization purposes illustrate thisenvironmental pressure, which results, among other things, in an increase in discharges of variouscontaminants into environmental compartments, particularly rivers. Therefore, part of these dischargescan end up in suspended particulate matter, which is then considered as storage wells, which transit inrivers. River development, particularly the construction of dams, encourages the sedimentation of these contaminated particles over time. These sediments of anthropogenic origin, also called legacy sediments, are therefore witnesses to human activities and make it possible to reconstruct the temporal trajectories of contamination within watersheds. The Eure River, a major tributary of the Seine estuary, has experienced significant anthropogenic pressures since the twentieth century. The temporal reconstruction of anthropogenic pressures has required the combination of different methodological approaches: (i) a diachronic analysis of the morphological modifications of the river was carried out, in conjunction with (ii) an analysis of the sedimentary dynamics and the nature of the sediment deposits by coupling geophysical, sedimentological and geochemical methods, and (iii) the setting up of a network for monitoring the hydro-sedimentary behaviour with continuous sampling of suspended particulate matter. Significant geomorphological changes have occurred in the lower reaches of the watershed, with the main consequences being an outlet moved some ten kilometres in the direction of a dam and the formation of hydraulic annexes favouring the accumulation of sediments as early as the 1940s. These made it possible to show that the Eure River watershed had experienced significant contamination, the consequences of which are still being recorded despite the cessation of activities or uses. The temporal trends of trace metal and metalloid elements showed strong contaminations in As in the 1940s and contaminations of industrial origin in Cr, Co, Ni, Cu, Zn, Ag and Cd in the 1960s and 1970s, as well as contaminations in Sb and Pb in 1990–2000. The latter are still recorded despite the cessation of the activities responsible for the discharges, as evidenced by the results from the suspended particulate matter currently collected in the river. Like most trace metals, organic contaminants such as PAHs showed significant contamination during the 1940–1960s, with signatures indicating a predominantly pyrogenic origin. PCBs showed significant contamination during the period 1950–1970, in connection with the production and national uses of mixtures composed mainly of low chlorinated congeners. Finally, interest in a third family of persistent organic contaminants, organochlorine pesticides, showed the use of lindane and DDT, particularly during the 1940–1970 period, and highlighted the post-ban use of lindane and the presence of a metabolite of DDT several decades after the cessation of its use, in connection with the increase in erosion of cultivated soils

Conditional source-term estimation with laminar flamelet decomposition has been utilized to model the mean chemical source term in a predictive RANS simulation of two different problems. With this model, integral equations relating the unconditional mean temperature and species fields to the conditional means are used to determine the flame structure as a combination of basis functions formed from solutions to the unsteady laminar flamelet equations. In the simulation of the Sandia Flame 'D', a well studied co-flowing piloted jet flame with a steady average solution, a converged solution was obtained which captured the trends in the temperature and major species, although the nitric oxide prediction overestimated the peak concentration by a substantial margin. Simulation of the autoignition process of non-premixed methane produced simulation predictions in excellent agreement with the experimental data. Using a new, more stringent, criteria to define ignition than earlier studies, the effect of ambient temperature on the ignition delay was captured, as was the expected physical behaviour prior to ignition. The errors in the simulation of the co-flowing piloted jet can be attributed to a large degree to the lack of ability of the basis functions used to account for the effect of the pilot flame. That the autoignition simulation was much more successful highlights the importance of including all relevant physics in the library of basis functions. Earlier formulations of the model were extended by expanding the sample space for the basis functions to include a wide variety of solutions to the laminar flamelet equations; in order to distinguish between the larger set, a new method of stabilizing the numerics was found to be necessary and the number of scalars used to determine the flame structure was increased. The formulation is easily extensible to libraries of basis functions which are capable of including effects such as pilot and edge flames not captured by the laminar flamelet solutions.

'Implementation of Conditional Source-term Estimation for the Prediction of Methane Ignition' The Laminar Flamelet Decomposition approach to Conditional Source-term Estimation is a recent development in combustion modeling that approximates the conditional averages of scalars as linear combinations of basis functions obtained from a library of unsteady solutions to the classical flamelet equations. Then an integral equation for a single scalar is inverted and solved for the weighting function of the basis functions. This weighting function is then used to obtain closure for the chemical source terms in the transport equations. Many parameters affect the solution of the Conditional Source-term Estimation method including the solution technique for inverting the integral equation; the number of solutions in the flamelet library; and the scalar used for the integral equation. The first part of this research investigated these parameters to optimize the solution for the Conditional Source-term Estimation. This is accomplished in the a priori sense by comparison to Direct Numerical Simulations of a simple turbulent flame. The second part of this work implemented Conditional Source-term Estimation into the commercially available multi-dimensional combustion software, KIVA-II by replacing the existing ignition mechanism. The model was used to predict the ignition of methane for different initial temperatures and compared to experimental results. The predicted ignition delay times were shorter than the experimental results but exhibited similar trends.