Thesis (MEng)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: In modern steelmaking a tundish serves as an important metallurgical reactor to remove inclusions and maintain thermal and chemical homogeneity in the product. In this study the flow behaviour in a four strand tundish was investigated by means of a 1/2 scale water model, based on Froude number similarity, as well as by using numerical modelling. Both the numerical study and physical model were used to characterise residence time distribution (RTD) in the vessel and to calculate properties pertaining to the tundish flow regime. The three different tundish configurations investigated in this study are: a bare tundish with no flow control devices, a tundish with a turbulence inhibitor and a tundish using a turbulence inhibitor with holes in combination with dams. Preliminary investigations focussed on the framework for obtaining an accurate numerical solution within reasonable computational times. The effect of assuming symmetry and dynamically steady flow in the numerical model was shown to be small relative to the effect of grid size and justifiable by the savings in computational time. The grid independence study indicated the importance of using a finer mesh in areas of high velocity gradients to obtain realistic results and also to limit the number of computational cells. A procedure using gradient adaptation was used to refine the mesh automatically in the required regions for different tundish geometries. Results also showed that the inlet boundary of the numerical model should be selected at the ladle outlet, since assuming a flat velocity profile at the nozzle port resulted in significant changes in the RTD response. Comparison of the results obtained using the numerical model with those from physical experiments yielded an average error of less than 10%. This was assumed to be a good prediction, considering the assumptions employed in the numerical model. Both the physical and numerical models showed that a tundish without flow control devices was prone to significant short circuiting. The addition of a turbulence inhibitor was shown to be successful in preventing short circuiting and provided surface directed flow, which is thought to aid inclusion removal in the slag. Additionally, the minimum, peak and mean residence times and plug flow volume fraction were increased significantly, while the dead volume fraction decreased. However, using a turbulence inhibitor with holes in combination with dams showed that this configuration may cause increased refractory wear together with increased risk of slag entrainment due to flow patterns with increased surface turbulence. It also showed that the short-circuiting might not be eliminated completely. This indicates that certain design changes to tundish flow control systems can introduce problems that outweigh the benefits of the altered flow patterns. Furthermore, the numerical method, which was based on the water model, was modified to simulate the high temperature steel process. A very good match was obtained between the results using the two different numerical models. This serves as additional evidence that tundish water modelling based on Froude number similarity provides a good representation of the actual industrial process. Using the numerical model based on the high temperature steel process the effect of turbulence inhibitor shape was studied for four different turbulence inhibitor designs. Results showed the best performance, based on flow characteristic properties and surface turbulence values, was achieved for the design using a rectangular box-like shape with flanges at the top. However, the comparison emphasized the effect of the turbulence inhibitor shape on the flow behaviour, as each design yielded completely different flow patterns. It was also observed that a good turbulence inhibitor provided an optimum amount of turbulent suppression. Insufficient suppression would cause fast flows, which will result in insufficient residence time for inclusion flotation and high surface turbulence values, which may cause slag entrainment. On the other hand, too much suppression may increase the variation between strands. / AFRIKAANSE OPSOMMING: Die verdeeltrog speel ‘n belangrike rol in die moderne staalvervaardigingsproses deur inklusies te verwyder en termiese en chemiese homogeniteit in die produk te verseker. In hierdie studie is die vloeigedrag in ‘n verdeeltrog met vier uitlate bestudeer deur middel van ‘n 1:2-skaal watermodel, gebaseer op Froude-getal gelykheid, sowel as deur numeriese modellering. Beide die numeriese en watermodel is gebruik om die verblyftyd-distribusie in die trog te karakteriseer en om waardes te bereken wat die vloeigedrag in die verdeeltrog beskryf. Drie verskillende verdeeltrog-opstellings is in hierdie studie bestudeer, naamlik: ‘n leë verdeeltrog met geen vloeibeheertoestelle nie, ‘n verdeeltrog met ‘n turbulensie- inhibeerder en ‘n verdeeltrog wat gebruik maak van ‘n turbulensie-inhibeerder met gate, gekombineer met lae damwalle. Vroeë ondersoeke het gefokus op ‘n metode om akkurate numeriese resultate binne aanvaarbare tye te verkry. Die invloed van die aannames van simmetrie en dinamiese gestadigde vloei op die resultate is bepaal om klein te wees teenoor dié van die roostergrootte. Die gebruik van die aannames is dus geregverdig deur die afname in berekeningstyd wat dit meebring. Die roosteronafhanklikheidstudie het getoon dat dit belangrik is om die vloeivergelykings oor klein volume-eenhede op te los in areas van hoë snelheidsgradiënte, beide om realistiese resultate te verseker en om die aantal volume-eenhede te beperk. ‘n Prosedure wat gebruik maak van gradiëntaanpassing is gebruik om die roostergrootte outomaties te verklein in die areas met hoë snelheidsgradiënte vir verskillende verdeeltrog-opstellings. Resultate het ook getoon dat dit belangrik is om die inlaatgrens van die numeriese model by die smeltpot se uitlaat te kies, aangesien noemenswaardige verskille in die verblyftyd-distribusie waargeneem is wanneer ‘n uniforme snelheidsprofiel aanvaar is waar die metaal die verdeeltrog binnegaan. Daar is gevind dat die resultate by die numeriese en watermodelle verskil met ‘n gemiddelde fout van ongeveer 10%. Dit word beskou as ‘n goeie ooreenstemming, as die aannames wat in die numeriese model gebruik word in ag geneem word. Beide die numeriese en watermodelle het getoon dat die metaalvloei in die leë verdeeltrog geneig is om die kortste roete na die uitlate te kies. Deur van ‘n turbulensie inhibeerder gebruik te maak word hierdie probleem opgelos deur die metaalvloei opwaarts te forseer. In die proses neem die minimum-, piek- en gemiddelde verblyftye, sowel as die propvloei volumefraksie, toe, terwyl die dooievloei fraksie afneem. Die turbulensie inhibeerder met die gate en die lae damwalle is egter meer geneig tot erosie van die vuurvaste materiaal en kan ook vloeipatrone veroorsaak wat metaalskuimdruppels in die metaal kan opneem. Verder volg ‘n klein fraksie van die vloei steeds ‘n kortpad na die uitlaat. Hierdie resultate toon dat sekere veranderings aan verdeeltrog vloeibeerheerstelsels nadelige gevolge kan hê wat enige positiewe gevolge uitkanselleer. Verder is die numeriese metode wat tot dusver op die watersisteem gebaseer is, aangepas om die vloei in die hoëtemperatuur staalproses te simuleer. ‘n Baie goeie vergelyking is verkry tussen die resultate wat die twee numeriese modelle voorspel. Hierdie dien as ‘n verdere bewys dat ‘n watermodel, wat op Froude-getal gelykheid gebaseer is, die industriële proses akkuraat kan simuleer. Deur gebruik te maak van die numeriese model van die hoë temperatuur staalproses, is die effek van die turbulensie-inhibeerder se vorm vir vier verskillende ontwerpe bestudeer. Die studie het getoon dat die beste resultate vir vloeikarakteriserende eienskappe en oppervlak- turbulensie verkry is met ‘n reghoekige ontwerp. Die resultate beklemtoon egter die belangrikheid van die invloed van die turbulensie-inhibeerder se vorm op die vloeigedrag in die verdeeltrog, aangesien elke vorm noemenswaardige verskillende vloeipatrone opgelewer het. Daar is opgemerk dat ‘n goeie turbulensie-inhibeerder ‘n optimale hoeveelheid turbulensie onderdrukking veroorsaak. In die geval van te min onderdrukking is die verblyftye te kort en die oppervlak turbulensie te hoog. Te veel onderdrukking kan egter lei tot groot variasie in die eienskappe van die metaal by die verskillende uitlate.

Climate change is predicted to have a global effect on temperatures and precipitation rates throughout the world. The UK Climate projections expect that in the United Kingdom this will lead to warmer, drier summers and wetter winters, where events of extreme rainfall are more common. These changes are expected to impact on slope hydrology, and concurrently slope stability. In the United Kingdom this impact is expected to be negative, whereas in other countries, such as Italy and France it could lead to slopes being more stable. Infrastructure slopes in the UK range in age and construction quality, they are susceptible to serviceability problems, characterised by heterogeneous material properties and can fail unexpectedly due to progressive reduction in soil shear strength. In this thesis the effects of climate change on a highway cutting in the south of England are modelled, using numerical methods. A finite element model is created and developed in the software package GeoStudio VADOSE/W. The model has been validated against observed pore water pressure trends and magnitudes and is shown to be able to accurately replicate the behaviour. By incorporating the effects of desiccation cracking on the soil s material properties, by the means of bimodal soil water characteristic curve and hydraulic conductivity function, the replication of these trends is improved even further. A series of future climate series were created using the UKCP09 Weather Generator 2.0. These series were implemented with the VADOSE/W model as climate boundary conditions and models were run, and the results compared to control, current climate results. The results were investigated by the means of statistical analyses which revealed that climate change will have some significant effects on the slope s hydrology, increasing magnitudes of evapotranspiration greatly which can have further significant effects on the magnitude of suctions developing in the slope throughout the summer. It is thought that the results suggest that climate change will not have significant negative effects on slope stability. However it is important to remember that the results only apply with certainty to the specific slope and climate change scenario investigated here. The methods used and developed within this thesis can be extended to other locations, in the UK and internationally, analysing the effects of different climate change scenarios.

Thesis (MSc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: The investigation of the interrelationship of the fluvial morphology and the salinity of the Great Fish River Estuary was performed by the combination of a two-dimensional morphological model and a one dimensional advection dispersion module. Two scenarios were defined for investigation, namely Scenarios A and B. Model bathymetry and grid/network for each model and scenario was compiled from topographical information obtained from aerial photos, SRTM data, LIDAR and 24 measured river cross sections of the area from the river mouth up to 27km upstream of the river mouth. Model boundary conditions were developed from empirical formulas and measured data from the Department of Water and Sanitation (DWS). Both models were calibrated with results obtained during field measurement conducted from 5 – 7 May 2012. Scenario A consisted of a long term 5 year morphological simulation (1 May 2007 to 30 May 2012) with manual mouth closure events for identified river low flow periods. Water levels upstream of the river mouth were extracted from the two-dimensional morphological module and used as the downstream boundary condition of the one-dimensional advection dispersion (salinity) model. For scenario B floods with return periods between 1:2 and 1:100 years were simulated in the morphological model. The resultant bathymetries were then used to compile the network and bathymetry of the one dimensional advection dispersion (salinity) model. The different flood resultant bathymetries were then used with equal boundary conditions (representative of the average flow in the river) in the one dimensional salinity model. The predicted salinity was compared for each bathymetry used. From model results distinct trends were observed. During low flow conditions the estuary basin fills with sediments and during floods the sediments are flushed out of the estuary. Large magnitude floods greatly erode the estuary especially in the middle reach, during floods the tidal inlet experiences overtopping and subsequent erosion, the constriction at the tidal inlet is completely destroyed during larger floods. The estuary mouth in its closed state experiences slight overtopping and the mouth is breached during periods of high river flows. The magnitude of salt intrusion depends mainly on the size (the constriction) of the river mouth (tidal inlet). During periods of mouth closure the average salinity in the estuary decreases, average salinity increases if the tidal inlet area is increased. The extent of salt intrusion is approximately 10 km upstream of the river mouth when the mouth is open and the intrusion length increases during spring tides. / AFRIKAANSE OPSOMMING: Die verwantskap tussen die Groot Vis Rivier Estuarium morfologie en sout toestand is ondersoek deur die kombinasie van ‘n twee dimensionele morfologiese model en ‘n een dimensionele sout model. Twee toestande is gedefinieer vir ondersoek naamlik Scenario A en Scenario B. Die area vanaf die rivier mond tot 27 km stroomop van die rivier mond is deur die modelle gesimuleer. Die gemodeleerde area stem ooreen met die area waar gemete rivier-snit diepte metings beskikbaar was, onbrekende data is aangevul met behulp van lugfotos, LIDAR- en SRTM- data. Die model grens toestande is bepaal met empiriese vereglykings asook gemete data vanaf die Departement Waterwese. Beide numeriese modelle was gekalibreer met veld data verkry vanaf 5 tot 7 Mei 2012. Scenario A het behels ‘n langtermyn 5 jaar morfologiese modellering (1 Mei 2007 tot 30 Mei 2012) met toemond toestande gedurende gedefinieerde rivier vloei toestande. Die watervlakke van die twee dimensionele morfologiese model stroomop van die mond is gebruik as die stroomaf grenstoestand van die een dimnesionele model om die effek van die morfologiese veranderinge te inkorporeer in die sout model. Scenario B het behels die simulering van rivier vloede met herhaal periodes tussen 1:2 en 1:100 jaar in die morfologiese model. Die rivier-bodem vlakke verkry van laasgenoemde simulasiesis toe gebruik in die een dimensionele sout model met dieselfde grenstoestande wat ooreenstem met die gemiddelde toestande in die Groot Vis Rivier. Aangesien die grenstoestande dieselfde was en net die rivier-bodem vlakke gevarieer is, kon die effek van vloede op die souttoestand in die estuarium bepaal word. Uit die model resultate kon duidelike tendense waargeneem word. Gedurende lae rivier vloei toestande is die estuarium gevul met sediment en tydens vloede het die sediment weer ge-erodeer en gedeponeer in die oseaan. Groot vloede veroorsaak baie erosie in die estuarium veral in die middel bereik en by die riviermonding. Indien die vloed groot genoeg is word die riviermond vernouing totaal uitgespoel. Die riviermonding in die geslote staat ondervind effense oorstroming en word oopgespoel indien die rivier vloei groot genoeg is. Die graad van die sout indringing in die estuarium hang hoofsaaklik af van die grootte (die vernouing) van die rivier mond. Gedurende toe mond toestande is die gemiddelde sout vlakke in die estuarium laer, wanner die rivier monding groter raak, word die gemiddelde sout vlakke in die estuarium meer. Die omvang van die sout indringing strek tot ongeveer 10 km stroomop van die rivier mond wanneer die mond oop is en die indringing afstand neem toe gedurende spring getye.

This thesis is based on the PhD work of investigating the Friction Stir Welding process (FSW) with numerical and Artificial Neural Network (ANN) modelling methods. FSW was developed at TWI in 1991. As a relatively new technology it has great advantages in welding aluminium alloys which are difficult to weld with traditional welding processes. The aim of this thesis was the development of new modelling techniques to predict the thermal and deformation behaviour. To achieve this aim, a group of Gleeble experiments was conducted on 6082 and 7449 aluminium alloys, to investigate the material constitutive behaviour under high strainrate, near solidus conditions, which are similar to what the material experiences during the FSW process. By numerically processing the experimental data, new material constitutive constants were found for both alloys and used for the subsequent FSW modelling work. Importantly no significant softening was observed prior to the solidus temperature. One of the main problems with numerical modelling is determining the values of adjustable parameters in the model. Two common adjustable parameters are the heat input and the coefficients that describe the heat loss to the backing bar. To predict these coefficients more efficiently a hybrid model was created which involved linking a conventional numerical model to an ANN model. The ANN was trained using data from the numerical model. Then thermal profiles were abstracted (summarised) and used as inputs; and the adjustable parameters were used as outputs. The trained ANN could then use abstracted thermal profiles from welding experiments to predict the adjustable parameters in the model. The first stage involved developing a simplified FE thermal model which represents a typical welding process. It was used to find the coefficients that describe the heat loss to the backing bar, and the amount of power applied in the model. Five different thermal boundary conditions were studied, including both convective and ones that included the backing bar with a contact gap conductance. Three approaches for abstracting the thermal curves and using as inputs to the ANN were compared. In the study, the characteristics of the ANN model, such as the ANN topology and gradient descent method, were evaluated for each boundary condition for understanding of their influences to the prediction. The outcomes of the study showed that the hybrid model technique was able to determine the adjustable parameters in the model effectively, although the accuracy depended on several factors. One of the most significant effects was the complexity of the boundary condition. While a single factor boundary condition (e.g. constant convective heat loss) could be predicted easily, the boundary condition with two factors proved more difficult. The method for inputting the data into the ANN had a significant effect on the hybrid model performance. A small number of inputs could be used for the single factor boundary condition, while two factors boundary conditions needed more inputs. The influences from the characteristics of the ANN model were smaller, but again thermal model with simpler boundary condition required a less complex ANN model to achieve an accurate prediction, while models with more complex boundary conditions would need a more sophisticated ANN model. In the next chapter, the hybrid method was applied to a FSW process model developed for the Flexi-stir FSW machine. This machine has been used to analyse the complex phase changes that occur during FSW with synchrotron radiation. This unique machine had a complex backing bar system involving heat transfer from the aluminium alloy workpiece to the copper and steel backing bars. A temperature dependent contact gap conductance which also depends on the material interface type was used. During the investigation, the ANN model topologies (i.e. GFF and MFF) were studied to find the most effective one. Different abstracting methods for the thermal curves were also compared to explore which factors (e.g. the peak temperature in the curve, cooling slope of a curve) were more important to be used as an input. According to close matching between the simulation and experimental thermal profiles, the hybrid model can predict both the power and thermal boundary condition between the workpiece and backing bar. The hybrid model was applied to six different travel speeds, hence six sets of heat input and boundary condition factors were found. A universal set was calculated from the six outcomes and a link was discovered between the accuracy of the temperature predictions and the plunge depth for the welds. Finally a model with a slip contact condition between the tool and workpiece was used to investigate how the material flow behaviour was affected by the slip boundary condition. This work involved aluminium alloys 6082-T6 and 7449-T7, which have very different mechanical properties. The application of slip boundary condition was found to significantly reduce the strain-rate, compared to a stick condition. The slip condition was applied to the Flexi-stir FSW experiments, and the results indicated that a larger deformation region may form with the slip boundary condition. The thesis successfully demonstrates a new methodology for determining the adjustable parameters in a process model; improved understanding of the effect of slip boundary conditions on the flow behaviour during FSW and insight in to the behaviour of aluminium alloys at temperatures approaching the solidus and high strain-rates.

Deformation in continents is not restricted to narrow bands but is spread over great distances within their interiors. A number of lines of evidence, including the distribution of earthquakes, reveal that the strength of different continental regions varies markedly. While it is relatively easy to qualitatively map out these variations, little progress has been made in quantifying the range of strength in the continents and identifying the physical mechanisms that control these variations. I investigate crustal deformation associated with the earthquake cycle, inflation of magma chambers beneath volcanoes, and changes in surface loads. Results of these models has important implications for our understanding of large-scale continental deformation and mountain building, in addition to both seismic and volcanic hazard assessment. Novel analytic solutions for simple shear with depth-dependent linear and non-linear viscoelastic rheologies are derived, in addition to analytical solutions for imposed harmonic tractions and displacements on an elastic layer over a Maxwell viscoelastic half space.

The Aquifer Thermal energy Storage (ATES) system is one of the most important techniques, that use the aquifer as an energy storage for heating and cooling the buildings. The operation of the ATES system is based on pumping the cold ground water into the buildings during the summer and then injecting the warmer water back into the aquifer. During the winter, the direction is reversed. Warm water is pumped to the buildings, where it serves for the heating system and then it is injected back into the aquifer. The ATES system has a main effect on the chemical composition of ground water, because of changes in the temperature of ground water, mixing of water from different redox zone and increasing of ground water flow in the neighbourhood of the wells of the ATES system. The modification in chemical composition and temperature could influence the behaviour of contaminants. If we want to install the ATES system into a contaminated aquifer, we have to know, if the ATES system would influence the behaviour of the contaminant plume positively or negatively. This study was made to increase the understanding of the dependence between changes of temperature and oxidation-reduction conditions in the ATES system and biodegradation of chlorinated hydrocarbons. In the first part, several scenarios were modelled...

Ce manuscrit s'intéresse au devenir de contaminants organiques dans le sol, et plus précisément celui des composés organo-halogénés volatils (COHV).Il propose des outils d'aide à la décision en utilisant la modélisation numérique appliquée à des problématiques environnementales portant sur le traitement de sites et sols pollués par des COHV. Il présente, à ma connaissance, la première modélisation numérique en 3D, grâce au simulateur TMVOC, d'une technique de dépollution physique, à savoir le pompage réussi au sein d'un aquifère d'un liquide en phase non-aqueuse plus dense que l'eau ou Dense Non-Aqueous Phase Liquid (DNAPL). Les très bons résultats de cette simulation permettent d’envisager l’optimisation d’un système de pompage asservi pour dépolluer un site contaminé aux COHV. Cette thèse s'intéresse aussi à une méthode de d'évaluation, à la fois qualitative et quantitative, de l'efficacité du pompage : un test de traçage utilisant des traceurs bisolubles à coefficients de partage variables (partitioning interwell tracer test – PITT). Ce PITT permet de connaître à la fois la répartition spatiale, au sein d’un aquifère, d’une bulle de DNAPL et aussi d’en évaluer sa saturation et donc son volume. Ce manuscrit met à disposition des outils élaborés et adaptés au traitement d'un DNAPL dans un aquifère. La combinaison des deux techniques présentées, à savoir la modélisation d'un pompage de DNAPL et un PITT, sont parfaitement reproductibles dans des environnements similaires et à l'échelle industrielle. Enfin, ces méthodes permettent de réduire considérablement les coûts de caractérisation (PITT) et d'exploitation par l'optimisation de systèmes de pompage / This manuscript deals with the fate of organic contaminants in soil, more precisely of volatile organo-chlorinated compounds (VOHC) and offers some decision making techniques and tools using numerical modelling applied to environmental issues about the treatment of soils contaminated by VOHC. It presents, to the best knowledge of the auhor, the first 3D numerical modelling, with the simulator TMVOC, of a physical treatment technique, namely the successful pumping within an aquifer, of a Dense Non-Aqueous Phase Liquid (DNAPL). The very good results of this simulation give the opportunity to design a controlled and automated pumping system to remediate a polluted site. This thesis also deols with a tracer test assessment method, both qualitative and quantitative, for the efficiency of this pumping : a partitioning interwell tracer test (PITT). The PITT allows us to determine the spatial repartion of the DNAPL and also to assess its saturation, hence its volume. This manuscript offers elaborated tools adapted to the remediation of a DNAPL in an aquifer. The combination of these two techniques, namely the DNAPL pumping and the PITT, are perfectly reproducible in similar environments up to an industrial scale. Finally, exploitation and characterisation costs for DNAPL remediation can be extremely reduced by numerical modelling and optimisation

L'approche multi-échelle FEMxDEM est une méthode numérique innovante pour les problèmes géotechniques impliquant des matériaux granulaires. La méthode des éléments finis (FEM) et la méthode des éléments discrets (DEM) sont simultanément appliquées à résoudre, respectivement, le problème structurel à la macro-échelle et la microstructure du matériau à la micro-échelle. L'avantage d'utiliser une telle configuration à double échelle est de permettre d'étudier un problème d'ingénierie sans la nécessité de lois de comportement standard, capturant ainsi l'essence des propriétés des matériaux. Le lien entre les échelles est obtenu par homogénéisation numérique, de sorte que la loi de comportement continu numérique et la matrice tangente correspondante sont obtenues directement à partir de la réponse discrète de la microstructure.En règle générale, l'approche FEMxDEM présente quelques inconvénients; la vitesse de convergence et la robustesse de la méthode ne sont pas aussi efficaces que dans les modèles FEM classiques. De plus, le coût de calcul de l'intégration de la micro-échelle et la dépendance du maillage typique de la macro-échelle, rendent l'approche multi-échelle FEMxDEM discutable pour des utilisations pratiques. Le but de ce travail est de se concentrer sur ces questions théoriques et numériques avec l'objectif de rendre l'approche multi-échelle FEMxDEM robuste et applicable à des configurations à l'échelle réelle. Une variété d'opérateurs est proposée afin d'améliorer la convergence et la solidité de la méthode dans un cadre quasi-Newton. L'indépendance de l’intégration des différents points de Gauss et les caractéristiques d’intensivité sur les d'éléments sont exploités par l'utilisation d’une parallélisation en utilisant un paradigme OpenMP. Au niveau macro, une relation constitutive second gradient est mise en œuvre afin d'enrichir la relation de Cauchy de premier gradient apportant indépendance du maillage au modèle.Les améliorations susmentionnées rendent l'approche FEMxDEM compétitive avec les modèles FEM classiques en termes de coût de calcul permettant ainsi d'effectuer des simulations multi-échelle FEMxDEM robustes et indépendantes du maillage, depuis l'échelle du laboratoire (par exemple essaie biaxiale test) jusqu’à celle du problème à l'échelle de l'ingénierie (par exemple, excavation d’une galerie).Mots clés:Double échelle, homogénéisation numérique, loi constitutive numérique, élasto-plasticité, second gradient, matériaux microstructurés, grande déformation, éléments finis, éléments discrets, méthode de Newton, parallélisation, unicité. / The multi-scale FEMxDEM approach is an innovative numerical method for geotechnical problems involving granular materials. The Finite Element Method (FEM) and the Discrete Element Method (DEM) are simultaneously applied to solve, respectively, the structural problem at the macro-scale and the material microstructure at the micro-scale. The advantage of using such a double scale configuration is that it allows to study an engineering problem without the need of standard constitutive laws, thus capturing the essence of the material properties. The link between scales is obtained via numerical homogenization, so that, the continuum numerical constitutive law and the corresponding tangent matrix are obtained directly from the discrete response of the microstructure.Typically, the FEMxDEM approach presents some drawbacks; the convergence velocity and robustness of the method are not as efficient as in classical FEM models. Furthermore, the computational cost of the microscale integration and the typical mesh-dependency at the macro-scale, make the multi-scale FEMxDEM approach questionable for practical uses. The aim of this work is to focus on these theoretical and numerical issues with the objective of making the multiscale FEMxDEM approach robust and applicable to real-scale configurations. A variety of operators is proposed in order to improve the convergence and robustness of the method in a quasi-Newton framework. The independence of the Gauss point integrations and the element intensive characteristics of the code are exploited by the use of parallelization using an OpenMP paradigm. At the macro level, a second gradient constitutive relation is implemented in order to enrich the first gradient Cauchy relation bringing mesh-independency to the model.The aforementioned improvements makes the FEMxDEM approach competitive with classical FEM models in terms of computational cost thus allowing to perform robust and mesh-independent multi-scale FEMxDEM simulations, from the laboratory scale (e.g. biaxial test) to the engineering-scale problem, (e.g. gallery excavation).Keywords:Double scale, numerical homogenization, numerical constitutive law, elasto-plasticity, second gradient, microstructured materials, large strain, finite elements, discrete elements, Newton method, parallelization, uniqueness.

Seawater intrusion (SWI) is a widespread environmental problem, particularly in arid and semi-arid coastal areas. Unplanned prolonged over-pumping of groundwater is the most important factor in SWI that could result in severe deterioration of groundwater quality. Therefore, appropriate management strategies should be implemented in coastal aquifers to control SWI with acceptable limits of economic and environmental costs. This PhD project presents the development and application of a simulation-optimization (S/O) model to assess different management methods of controlling saltwater intrusion while satisfying water demands, and with acceptable limits of economic and environmental costs, in confined and unconfined coastal aquifers. The first S/O model (FE-GA) is developed by direct linking of an FE simulation model with a multi-objective Genetic Algorithm (GA) to optimize the efficiency of a wide range of SWI management scenarios. However, in this S/O framework, several multiple calls of the simulation model by the population-based optimization model, evaluating best individual candidate solutions resulted in a considerable computational burden. To solve this problem the numerical simulation model is replaced by an Evolutionary Polynomial Regression (EPR)-based surrogate model in the next S/O model (EPR-GA). Through these S/O approaches (FE-GA and EPR-GA) the optimal coordinates and rates of the both abstraction and recharge barriers are determined in the studied management scenarios. As a result, a new combined methodology, so far called ADRTWW, is proposed to control SWI. The ADRTWW model consists of deep Abstraction of saline water near the coast followed by Desalination of the abstracted water to a potable level for public uses and simultaneously Recharging the aquifer using a more economic source of water such as treated wastewater (TWW). In accordance to the available recharge options (injection through well or infiltration from surface pond), the general performance of ADRTWW is evaluated in different hydro-geological settings of the aquifers indicating that it offers the least cost and least salinity in comparison with other scenarios. The great capabilities of both developed S/O models in identification of the best management solutions and the optimal coordinates and rates of the abstraction well and recharge well/pond are discussed. Both FE-GA and EPR-GA can be successfully employed by a robust decision support system. In the next phase of the study, the general impacts of sea level rise (SLR), associated with its transgression nature along the coastline surface on the saltwater intrusion mechanism are investigated in different hypothetical and real case studies of coastal aquifer systems. The results show that the rate and the amount of SWI are considerably greater in aquifers with flat shoreline slopes compared with those with steep slopes. The SWI process is followed by a significant depletion in quantity of freshwater resources at the end of the century. The situation is exacerbated with combined action of SLR and groundwater withdrawals. This finding is also confirmed by 3D simulation of SWI in a regional coastal aquifer (Wadi Ham aquifer) in the UAE subjected to the coupled actions of SLR and pumping.

No description available.