Among all casting defects, shrinkage porosities could significantly reduce the strength of metal parts. As several critical components in aerospace and automotive industries are manufactured through casting processes, ensuring these parts are free of defects and are structurally sound is an important issue. This study investigates the formation of shrinkage-related defects in alloy solidification. To have a better understanding about the defect formation mechanisms, three sets of experimental studies were performed. In the first experiment, a real-time video radiography technique is used for the observation of pore nucleation and growth in a wedge-shaped A356 aluminum casting. An image-processing technique is developed to quantify the amount of through-thickness porosity observed in the real-time radiographic video. Experimental results reveal that the formation of shrinkage porosity in castings has two stages: 1-surface sink formation and 2- internal porosity evolution. The transition from surface sink to internal porosity is defined by a critical coherency limit of . In the second and third experimental sets, two Manganese-Steel (Mn-Steel) castings with different geometries are selected. Several thermocouples are placed at different locations in the sand molds and castings to capture the cooling of different parts during solidification. At the end of solidification, castings are sectioned to observe the porosity distributions on the cut surfaces. To develop alloys’ thermo-physical properties, MAGMAsoft (a casting simulation software package) is used for the thermal simulations. To assure that the thermal simulations are accurate, the properties are adjusted to get a good agreement between simulated and measured temperatures by thermocouples. Based on the knowledge obtained from the experimental observations, a mathematical model is developed for the prediction of shrinkage porosity in castings. The model, called “advanced feeding model”, includes 3D multi-phase continuity, momentum and pore growth rate equations which inputs the material properties and transient temperature fields, and outputs the feeding velocity, liquid pressure and porosity distributions in castings. To solve the model equations, a computational code with a finite-volume approach is developed for the flow calculations. To validate the model, predicted results are compared with the experimental data. The comparison results show that the advanced feeding model can accurately predict the occurrence of shrinkage porosity defects in metal castings. Finally, the model is optimized by performing several parametric studies on the model variables.

In this study, piezometric heads in a composite aquifer composed of an alluvial deposit having a width adjacent to a semi-infinite fractured rock are determined. One dimensional transient flow induced by a constant discharge pumping rate from a stream intersecting alluvial part of the aquifer is considered. Parts of the aquifer are homogeneous andisotropic. Equations of flow, initial and boundary conditions are converted to dimensionless forms for graphical presentation and the interpretation of results independent of discharge and head inputs specific to the problem. Equations are solved first in Laplace domain and Laplace domain solutions are inverted numerically to real time domain by utilizing Stehfest algorithm.For this purpose, a set of subroutines in VBA Excel are developed. This procedure is verified by application of code to flow in semi-infinite homogeneous aquifer under constant discharge for which analytical solution is available in literature. VBA codes are also developed for two special cases of finite aquifer with impervious and with recharge boundary on the right hand side. Results of composite aquifer solutions with extreme tranmissivity values are compared with these two cases for verification of methodology and sensivity of results.

A naturally fractured reservoir (NFR) is a reservoir with a connected network of fractures created by natural processes such as diastrophism and volume shrinkage (Ordonez et al. 2001). There are two models to simulate this kind of reservoirs: the discrete fracture model and the dual porosity model. In the dual porosity model, the matrix blocks occupy the same physical space as the fracture network and are identical rectangular parallelepipeds with no direct communication between isotropic and homogeneous matrix blocks. However, each fracture and matrix property is defined separately in the discrete fracture model. Another feature of this thesis is tracer testing. In this process, a chemical or radioactive element is injected to the reservoirs, and then it can be traced using the devices, which are designed to detect the tracers. Tracer tests have several advantages such as determining residual oil saturation, identifying barriers or high permeability zones in reservoirs, and providing the information on flow patterns. Limited number of research studies has been done on performing tracer tests in naturally fractured reservoirs. Also because there is not enough information about the advantages and disadvantages of the discrete fracture and the dual porosity models, researchers and engineers lack the expertise to confidently select either the discrete fracture or the dual porosity models to simulate the different types of NFRs. In this thesis, we compared the oil and water productions, and tracer concentration curves in various reservoir conditions, using both the discrete fracture and the dual porosity models. We used the ECLIPSE, which is a commercial software package in the area of petroleum industry, to model a naturally fractured reservoir. We performed a simple waterflooding with two conservative tracers on the reservoirs. The results presented in each section include the graphs of the oil production rate, water production rate, and tracer concentration. In addition, we presented the oil saturation profiles of a cross-section, which includes the production and injection wells. The results illustrated that both the discrete fracture and the dual porosity models are in good agreement, except for a few special cases. Generally, the oil production using the dual porosity model is more than in the discrete fracture model. The major disadvantage of the dual porosity model is that the fluid distribution in the matrix blocks is changing homogenously during the waterflooding period. In other words, ECLIPSE shows a constant value of the oil and water saturations in each time step for the matrix blocks. However, the dual porosity model is 3 to 4 times faster than the discrete fracture model. In the discrete fracture model, the users have complete control in defining the reservoirs. For example, the fracture aperture, fracture spacing, and fracture porosities can be set by the user. The disadvantage of this model is that millions of grid blocks are needed to model a large reservoir with small fracture spacing. / text

Laplace transform step-response functions are presented for one dimensional transient flow in fractured semi-infinite &amp / finite aquifers. Unsteady flow in the aquifer resulting from a constant discharge pumped from the stream is considered. Flow is one-dimensional, perpendicular to the stream in the confined aquifers. The stream is assumed to penetrate the full thickness of the aquifer. The aquifers may be semi-infinite or finite in width. The Laplace domain solutions are numerically inverted to the real-time domain with the Stehfest (1970) algorithm. During the course of the thesis a simple computer code is written to handle the algorithm and the code is verified by applying it to the one-dimensional transient flow in a semi-infinite homogeneous aquifer problem which can be solved analytically to crosscheck with the numerical results.

Les réservoirs de type "tight oil" et "shale gaz" ont une partie importante de leur volume poreux occupée par des micropores (< 2nm) et des mesopores (entre 2 et 50 nm). Ce type d'environnement crée de fortes forces d’interaction dans le fluide confiné avec les parois du pores et entre ses propres molécules, ce qui change fortement la thermodynamique du fluide. Un travail important doit donc être effectué sur le développement de méthodes de mise à l'échelle de la distribution de pore pour effectuer des simulations réservoir à grande échelle. Premièrement, des simulations moléculaires sont effectuées sur des fluides confinés afin d'obtenir des propriétés thermodynamiques de référence à l'équilibre liquide/vapeur pour différentes tailles de pore. Ensuite, une comparaison des données de simulation moléculaire avec les résultats issus des équation d'état utilisées dans la littérature a permis de mettre en valeur la méthode de flash avec pression capillaire et changement du point critique comme la meilleure méthode existante pour décrire la physique du fluide confiné. Des simulations fines d'écoulement matrice/fracture ont donc été effectuées pour différentes tailles de pore. Des modèles de mise à l'échelle en maillage grossier ont été ensuite construits à partir du même cas synthétique et les résultats ont été comparés avec ceux des simulations de référence en maillage fin. Un nouveau modèle de triple porosité considérant fracture, petit pores et grand pores avec une approche MINC a donné des résultats très proches du maillage fin. Finalement un réservoir stimulé hydrauliquement à grande échelle a été simulé pour différentes distributions de pores avec le modèle développé. / Tight oil and shale gas reservoirs have a significant part of their pore volume occupied by micro (below 2nm) and mesopores (between 2 and 50nm). This kind of environment creates strong interaction forces in the confined fluid with pore walls as well as between its own molecules and then changes dramatically the fluid phase behavior. An important work has therefore to be done on developing upscaling methodology of the pore size distribution for large scale reservoir simulations. Firstly, molecular simulations are performed on different confined fluids in order to get reference thermodynamic properties at liquid/vapor equilibrium for different pore sizes. Then, the comparison with commonly used modified equation of state (EOS) in the literature highlighted the model of flash with capillary pressure and critical temperature and pressure shift as the best one to match reference molecular simulation results. Afterwards fine grid matrix/fracture simulations have been built and performed for different pore size distributions. Then, coarse grid upscaling models have then been performed on the same synthetic case and compared to the reference fine grid results. A new triple porosity model considering fracture, small pores and large pores with MINC (Multiple Interacting Continua) approach, has shown very good match with the reference fine grid results. Finally a large scale stimulated reservoir volume with different pore size distribution inside the matrix has been built using the upscaling method developed here.