Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-03-06T19:49:56Z No. of bitstreams: 1 johnnymoreiragomes.pdf: 4938794 bytes, checksum: fb03990a45c2c77e8ff60eae73a2d21d (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-03-06T20:27:17Z (GMT) No. of bitstreams: 1 johnnymoreiragomes.pdf: 4938794 bytes, checksum: fb03990a45c2c77e8ff60eae73a2d21d (MD5) / Made available in DSpace on 2017-03-06T20:27:17Z (GMT). No. of bitstreams: 1 johnnymoreiragomes.pdf: 4938794 bytes, checksum: fb03990a45c2c77e8ff60eae73a2d21d (MD5) Previous issue date: 2015-02-24 / FAPEMIG - Fundação de Amparo à Pesquisa do Estado de Minas Gerais / Este trabalho compara diferentes esquemas numéricos para a solução de modelos modernos para a eletrofisiologia de miócitos cardíacos. Apresentamos o Método de Uniformização - amplamente utilizado para a solução de problemas estocásticos em ciência da computação - e mostramos que, quando aplicado na resolução numérica de modelos cardíacos baseados em Cadeias de Markov de Tempo contínuo, aumenta substancialmente a estabilidade numérica em relação a métodos explícitos tradicionalmente utilizados, como o Método de Rush-Larsen e o Método de Euler Explícito. A formulação em Cadeias de Markov para estruturas subcelulares - como os canais iônicos - permite a descrição detalhada do comportamento elétrico de células cardíacas para importantes aplicações experimentais, como a simulação dos efeitos de drogas e toxinas sobre a atividade elétrica da membrana celular. No entanto, as equações diferenciais associadas às Cadeias de Markov para canais iônicos frequentemente trazem problemas de estabilidade numérica, que limitam fortemente o passo de tempo utilizado por esquemas explícitos. Com a utilização do Método de Uniformização foi possível aumentar significativamente a magnitude dos passos de tempo utilizados em simulações de três modelos da eletrofisiologia cardíaca baseados em Cadeias de Markov. Neste trabalho mostramos como é possível associar o Método de Uniformização a outros esquemas explícitos para a solução numérica de tais modelos, e como tais técnicas melhoram significativamente o desempenho computacional em relação a métodos explícitos tradicionais. Além disso, propomos extensões do método de Rush-Larsen e do método de Uniformização com segunda ordem de precisão para o desenvolvimento de esquemas explícitos de passo de tempo adaptativo, visando reduzir ainda mais o custo computacional em simulações com tolerância numérica estrita. / This work compares different numerical schemes for the solution of modern electrophysiology models for cardiac myocytes. We present the Uniformization Method - frequently applied to stochastic problems in computer science - which significantly increase the numerical stability when used for the solution of cardiac models based on Continuous Time Markov Chains, with respect to traditional explicit schemes such as Rush-Larsen Method and Foward Euler Method. The Markov Chains formulation for subcellular structures, e.g. ionic channels, enables an accurate description of the electrical behaviour of cardiac cells for important experimental applications, for instance the simulation of the effects of drugs or toxins on the electrical activity of the cell's membrane. However, the differential equations associated with the Markov Chains for ionic channels frequently cause problems of numerical stability, which severely limits the time step used by explicit schemes. By using the Uniformization Method we could significantly increase the time steps size in simulations of three models of cardiac electrophysiology based on Markov Chains. In this work we show how the Uniformization Method can be used along with other foward numerical schemes for the solution of these models, and how these techniques significantly improve the computational performance with respect to traditional numerical methods. In adition, we propose extensions of the Rush-Larsen method and the Uniformization method with second-order accuracy for developing foward time-adaptive techniques, aiming to reduce the computational cost of simulations with strict numerical tolerances.

Cette thèse vise à explorer la question du changement d'échelle temporelle en modélisation hydrologique conceptuelle. Les principaux objectifs sont : (i) étudier les effets du changement du pas de temps sur les performances, les paramètres et la structure des modèles hydrologiques ; (ii) mettre au point un modèle pluie-débit applicable à différents pas de temps. Notre point de départ est le modèle global journalier GR4J, développé à Irstea. Ce modèle a été choisi comme le modèle de référence à adapter à d'autres résolutions plus fines, jusqu'à des pas de temps infra-horaires, en suivant une approche descendante. Pour nos tests, nous avons construit une base de données de 240 bassins versants non influencés en France, à différents pas de temps allant de 6 minutes à 1 jour, en utilisant: (i) les données pluviométriques à 6 minutes et la réanalyse des lames d'eau journalières à plus haute résolution spatiale ; (ii) les données de température journalière pour le calcul de l'évapotranspiration potentielle ; (iii) les données hydrométriques à pas de temps variable. Nous avons étudié l'impact de la distribution temporelle des entrées sur les performances du modèle en se focalisant sur la simulation de crue, sur la base de 2400 événements. Ensuite, notre évaluation du modèle a porté sur l'analyse de la cohérence des flux internes du modèle à différents pas de temps, afin d'assurer une performance satisfaisante à travers un fonctionnement du modèle cohérent. Notre diagnostic du modèle nous a permis d'identifier une amélioration de la structure du modèle à différents pas de temps infra-journaliers basée sur la complexification de la composante d'interception du modèle. / This thesis aims at exploring the question of temporal scaling in lumped conceptual hydrological modelling. The main objectives of the thesis are to: (i) study the effects of varying the modelling time step on the performance, parameters and structure of hydrological models; (ii) develop a hydrological model operating at different time steps, from daily to sub-hourly, through a unified, robust and coherent modelling framework at different time scales. Our starting point is the chain of conceptual rainfall-runoff models called ‘GR’, developed at Irstea, and in particular the daily ‘GR4J’ lumped model. The GR4J model will be the baseline model to be effectively downscaled up to sub-hourly time steps following a top-down approach. An hourly adaptation of this model had already been proposed in previous research studies, but some questions on the optimality of the structure at sub-daily time steps were still open. This thesis builds on these previous studies on the hourly model and responds to the operational expectations of improving and adapting the model at multiple sub-daily and sub-hourly time steps, which is particularly interesting for flood forecasting applications. For our modelling tests, we built a database of 240 unregulated catchments in metropolitan France, at multiple time steps, from 6-minute to 1 day, using fine time step hydro-climatic datasets available: (i) 6-min rain gauges and higher spatial-density daily reanalysis data for precipitation; (ii) daily temperature data for potential evapotranspiration (making assumptions on sub-daily patterns); (iii) sub-hourly variable time step streamflow data. We investigated the impact of the inputs temporal distribution on model outputs and performance in a flood simulation perspective based on 2400 selected flood events. Then our model evaluation focused on the consistency of model internal fluxes at different time steps, in order to ensure obtaining a satisfactory model performance by a coherent model functioning at multiple time steps. Our model diagnosis led us to identify and test a significant improvement of the model structure at sub-daily time steps based on the complexification of the interception component of the model. Thus, we propose a new version of the model at multiple sub-daily time steps, with the addition of an interception store without extra free parameters. Our tests also confirm the suitability at multiple time steps of a modified groundwater exchange function proposed earlier, leading to overall improved model accuracy and coherence.

Articulated haulers are fundamental equipment to transport material. The load carrying structure on a hauler consists mainly of welded frames. During welding of the frames high residual stress will be introduced. These stresses may have a significant impact on the fatigue life of the frames. This is the reason for having good knowledge of the weld residual stresses. The finite element method was used to calculate the residual stress distributions in a butt weld and a T-join weld. Simulation of the welding process with thermal and mechanical analysis was prepared by means of welding GUI implemented in LS-PrePost. The welding simulation is a computer intensive operation with high CPU time. That is why it is important to investigate which process factors that have the largest impact on welding simulation results. The aim of this thesis is to investigate the correlation between designed models in FEA software with published results of weld residual stress measurements and conclude which parameters should be mainly taken into consideration.

This thesis deals with stochastic water demand modellling in the water supply network. In the opening section, a research is created, where two different approaches to stochastic modelling of water consumption are presented. The practical part describes the creation of a deterministic hydraulic model and its calibration. Generated stochastic water demand patterns with a small time step of 1 minute, are then inserted into this model. Each household is assigned with a unique water demand pattern. Then a hydraulic analysis was done. A comparison of deterministic and stochastic approaches is presented at the end of the thesis. The comparison shows, that small-time step modelling does not have a big impact on the pressure ratios in the water supply network, but has a huge impact on the maximum flows and speeds occurring in links of the hydraulic model.

This diploma thesis deals with problem of water hammer in pipes affected by viscoelastic behavior of pipe material. Mathematical model uses pressure dependent speed of sound in water air mixture. For purpose of numerical solution the Method of Characteristics with independent time step is introduced. This method is compared with commonly used methods such as Method of Characteristics (MOC) and Lax Wendroff scheme. Derived model, solved with Method of Characteristics with independent time step, is verified with experimental simulation.

The numerical simulation of a moving heat source from a fixed point observer is often done by discretely adjusting its position over the steps of a thermal transient analysis. The efficiency of these simulations is increased when using a coarse temporal discretization whilst maintaining the quality of results. One systematic error source is the rare update of a nonconstant moving heat source with regard to its magnitude and location. In this work, we present an analysis of the error and propose a correction approach based on conserving the specified heat from a continuous motion in analyses with large time-step sizes. Deficiencies associated with the correction in special motion situations are identified by means of performance studies and the approach is extended accordingly. The advantages of applying the proposed correction are demonstrated through examples.

On étudie dans le cadre de la thèse une famille de schémas numériques permettant de résoudre les équations de Saint-Venant. Ces schémas utilisent une décomposition d'opérateur de type Lagrange-projection afin de séparer les ondes de gravité et les ondes de transport. Un traitement implicite du système acoustique (relié aux ondes de gravité) permet aux schémas de rester stable avec de grands pas de temps. La correction des flux de pression rend possible l'obtention d'une solution approchée précise quel que soit le régime d'écoulement vis-à-vis du nombre de Froude. Une attention toute particulière est portée sur le traitement du terme source qui permet la prise en compte de l'influence de la topographie. On obtient notamment la propriété dite équilibre permettant de conserver exactement certains états stationnaires, appelés état du "lac au repos". Des versions 1D et 2D sur maillages non-structurés de ces méthodes ont été étudiées et implémentées dans un cadre volumes finis. Enfin, une extension vers des méthodes ordres élevés Galerkin discontinue a été proposée en 1D avec des limiteurs classiques ainsi que combinée avec une boucle MOOD de limitation a posteriori. / In this thesis we study a family of numerical schemes solving the shallow water equations system. These schemes use a Lagrange-projection like splitting operator technique in order to separate the gravity waves and the transport waves. An implicit-explicit treatment of the acoustic system (linked to the gravity waves) allows the schemes to stay stable with large time step. The correction of the pressure fluxes enables the obtain of a precise approximation solution whatever the regime flow is with respect to the Froude number. A particular attention has been paid over the source term treatment which permits to take the topography into account. We especially obtain the so-called well-balanced property giving the exact conservation of some steady states, namely the "lake at rest" state. 1D and 2D versions of this methods have been studied and implemented in the finite volumes framework. Finally, a high order discontinuous Galerkin extension has been proposed in 1D with classical limiters along with a combined MOOD loop a posteriori limiting strategy.

With steadily increasing computational power, computational fluid dynamics (CFD) can be applied to unsteady problems such as seakeeping simulations. Therefore, a good balance between accuracy and computational speed is required. This thesis investigates the application of CFD to seakeeping performance predictions and aims to propose a best-practice procedure for efficient seakeeping simulations. The widely used KVLCC2 research vessel serves as a test case for this thesis and FINEŠ/Marine software package is used for CFD computations. In order to validate the simulations, results are compared to recent experimental data from SSPA as well as predictions with potential ˛ow code SHIPFLOW® Motions. As for the calm water simulations, both inviscid and viscous ˛ow computations are performed in combination with three mesh refinement levels. Seakeeping simulations with regular head waves of different wavelengths are set-up correspondingly. Furthermore, different strategies for time discretization are investigated. With the given computational resources, it is not feasible to complete seakeeping simulations with a ˝ne mesh. However, already the coarse meshes give good agreement to experiments and SHIPFLOW® Motions' predictions. Viscous ˛ow simulations turn out to be more robust than Euler ˛ow computations and thus should be preferred. Regarding the time discretization, a fixed time discretization of 150 steps per wave period has shown the best balance between accuracy and speed. Based on these findings, a best-practice procedure for seakeeping performance predictions in FINEŠ/Marine is established. Taking the most efficient settings obtained from head wave simulations, the vessel is subjected to oblique waves with 160° encounter angle. Under similar wave conditions, CFD predictions of a similar thesis show close agreement in terms of added wave resistance. Compared to the previous head wave conditions of this study, added resistance in 160° oblique waves is found to be significantly higher. This underlines that oblique bow quartering waves represent a relevant case for determining the maximum required power of a ship. CFD and potential ˛ow show similar accuracy with respect to ship motions and added wave resistance, albeit potential ˛ow outperforms CFD in terms of computational speed. Hence, CFD should be applied in cases where viscous effects are known to have large influence on a vessel's seakeeping behavior. This can be the case if motion control and damping devices are to be evaluated, for instance. / Tack vare den stadigt ökande beräkningskraften kan beräkningsuiddynamik (CFD) idag användas på beräkningsintensiva problem som sjöegenskapssimulationer. Den här rapporten undersöker användning av CFD på sjöegenskapsprestanda och syftar till att foreslå ett best-practice förfaringssätt för effektiv sjöegenskapssimulationer. Forskningsskrovet KVLCC2 fungerar som ett testfall för denna rapport och FINE—/Marine-mjukvarupaketet används för CFD-beräkningar. Viktiga parametrar, såsom ödestyp, beräkningsnät och tidssteg varierars systematiskt. Resultaten jämförs med experiment gjorda vid SSPA. Baserat på resultaten förelås en best-practice. Den föreslagna best-practice användas vidare för berökningar av sjöegenskaper i sneda vågor. Jämförelse av resultaten med liknande studier visar god överensstämmelse. Genom att använda det föreslagna förfarandet för best-practice kan CFD-sjöegenskapssimulationer användas på fall där viskösa krafter måste beaktas, till exempel rörelseregleringsanordningar.

No description available.

Nous présentons dans cette thèse les évolutions apportées au solveur Galerkin Discontinu Teta-CLAC, issu de la collaboration IRMA-AxesSim, au cours du projet HOROCH (2015-2018). Ce solveur permet de résoudre les équations de Maxwell en 3D, en parallèle sur un grand nombre d'accélérateurs OpenCL. L'objectif du projet HOROCH était d'effectuer des simulations de grande envergure sur un modèle numérique complet de corps humain. Ce modèle comporte 24 millions de mailles hexaédriques pour des calculs dans la bande de fréquences des objets connectés allant de 1 à 3 GHz (Bluetooth). Les applications sont nombreuses : téléphonie et accessoires, sport (maillots connectés), médecine (sondes : gélules, patchs), etc. Les évolutions ainsi apportées comprennent, entre autres : l'optimisation des kernels OpenCL à destination des CPU dans le but d'utiliser au mieux les architectures hybrides ; l'expérimentation du runtime StarPU ; le design d'un schéma d'intégration à pas de temps local ; et bon nombre d'optimisations permettant au solveur de traiter des simulations de plusieurs millions de mailles. / In this thesis, we present the evolutions made to the Discontinuous Galerkin solver Teta-CLAC – resulting from the IRMA-AxesSim collaboration – during the HOROCH project (2015-2018). This solver allows to solve the Maxwell equations in 3D and in parallel on a large amount of OpenCL accelerators. The goal of the HOROCH project was to perform large-scale simulations on a complete digital human body model. This model is composed of 24 million hexahedral cells in order to perform calculations in the frequency band of connected objects going from 1 to 3 GHz (Bluetooth). The applications are numerous: telephony and accessories, sport (connected shirts), medicine (probes: capsules, patches), etc. The changes thus made include, among others: optimization of OpenCL kernels for CPUs in order to make the best use of hybrid architectures; StarPU runtime experimentation; the design of an integration scheme using local time steps; and many optimizations allowing the solver to process simulations of several millions of cells.