No description available.

Au cours de la production d'hydrocarbures, l'extraction de fluides fait décroître la pression de pore dans les réservoirs (« depletion »). Ceci induit un changement du champ de contraintes qui résulte en une augmentation des contraintes effectives appliquées sur le réservoir. Les mesures in situ montrent que les variations de contraintes peuvent être décrites par un paramètre appelé chemin de chargement (stress path), défini comme le rapport entre la variation de contrainte effective horizontale et la variation de contrainte effective verticale par rapport aux conditions initiales dans le réservoir. La compaction induite par la production d'hydrocarbures peut avoir de graves conséquences dans le cas de roches faiblement consolidées car elle induit des variations des propriétés pétrophysiques des roches in situ, notamment de la perméabilité, un des paramètres les plus importants pour estimer la performance d'un réservoir mais aussi un des plus difficiles à mesurer. Pour compliquer encore les choses, la perméabilité est souvent anisotrope dans les réservoirs avec de forts contrastes entre la perméabilité horizontale kh et la perméabilité verticale kv.L'objectif de cette étude est de comprendre l'influence des chemins de chargement sur le comportement mécanique et les évolutions couplées de perméabilité pour un sable quartzeux (Sable de la Durance, DS) et un grès faiblement consolidé (grès d'Otter Sherwood, OSS, qui constitue la roche réservoir du champ pétrolier de Wytch Farm en Angleterre). Nos résultats montrent que le grès peu consolidé présente un comportement mécanique similaire à celui de roches consolidées. Au contraire, le sable présente un comportement différent, avec une transition plus graduelle entre les régimes de déformation qui nécessite d'utiliser un critère basé sur l'évolution du rayon de courbure des courbes contraintes-déformations pour déterminer les contraintes limites : cette méthode a été validée par une étude d'analyse des émissions acoustiques pour caractériser l'endommagement. Les domaines de déformation élastique et plastique ont été bien définis et les contraintes limites ont été comparées aux prédictions du modèle Cam-Clay modifié et du modèle d'enveloppe limite normalisée. Les perméabilités horizontale et verticale ont été mesurées sous contraintes. Pour analyser l'influence des effets de bord dans les essais mécaniques, les perméabilités mesurées soit classiquement sur toute la longueur de l'échantillon, soit entre deux points intermédiaires ont été comparées. Pour l'écoulement horizontal, les facteurs géométriques et facteurs d'anisotropie ont été déterminés par des simulations numériques en éléments finis afin de pouvoir déterminer les vraies valeurs de perméabilité horizontale. L'évolution de la perméabilité suit l'évolution de la déformation des matériaux et est contrôlée aussi bien par la déformation volumique que par la déformation en cisaillement. A partir de nos mesures il est possible de séparer l'effet de la pression moyenne de l'effet de la contrainte déviatorique sur l'évolution de la perméabilité en construisant des cartes d'isoperméabilités dans l'espace des contraintes. Enfin une modélisation élasto-plastique a été réalisée pour prédire le comportement hydro-mécanique du grès faiblement consolidé. L'approche utilisée permet de prédire de manière satisfaisante l'évolution de la perméabilité avec les contraintes, à partir d'une loi exponentielle fonction de la déformation effective. Au contraire, pour le sable de la Durance le lien entre l'évolution de la perméabilité et la déformation est loin d'être évidente, notamment aux faibles contraintes où la réduction de perméabilité est très rapide. Pour mieux comprendre ces évolutions de perméabilité, une analyse de l'endommagement a été réalisée par des mesures sur échantillons et en utilisant des techniques d'imagerie à plusieurs échelles. / During hydrocarbon production, the extraction of fluid induces a decrease of pore pressure called depletion. This depletion causes a change in the stress field that results in an increased stress on the rock by enhancement of the effective stress in the reservoir. In situ measurements show that the stress variations can be described by the so-called stress path parameter, defined as the ratio of the change in effective horizontal stress by the change in effective overburden stress from initial reservoir conditions. This production induced compaction can have severe consequences in the case of poorly consolidated reservoirs. Compaction induces variations of petrophysical properties of in situ rocks and particularly permeability variations, one of the most important parameters controlling reservoir performance. Yet it is one of the most difficult property to measure. To complicate matters further, permeability anisotropy is often found in reservoirs. Therefore the horizontal permeability kh, may be different from the vertical permeability kv.The aim of this study is to understand the influence of stress paths on the mechanical behavior and coupled permeability evolutions of a Quartz sand (Durance Sand, DS) and a weakly consolidated sandstone (Otter Sherwood Sandtone, OSS which is the reservoir rock of the Wytch Farm oil field, UK). We found that the weakly consolidated rock presents a mechanical behavior similar to that of consolidated rocks. However, the sand shows a different behavior, with a gradual transition regime which requires the use of a curvature criterion to peak yield stresses on the stress-strain evolution plot; this criterion has been validated on the basis of Acoustic Emission analysis. The elastic and plastic deformation regimes are well identified and the determined yield stresses are fitted using the modified Cam-Clay and Elliptic Cap models for all observed onsets of plastic yielding. Both vertical and horizontal permeability have been measured during loading. To analyze the influence of end effects during loading in the triaxial cell, permeabilities measured over the mid-section and over the total core length were compared. For the horizontal flow, the geometrical and anisotropy factors were determined using Finite Element simulations in order to calculate the correct horizontal permeability. Permeability evolution follows closely the material deformation and is controlled by both volumetric and shear strains. It is possible to infer the effect of the mean pressure and/or the deviatoric stress on the permeability evolution by building isopermeability maps in the stress space. Finally, an application of elasto-plastic modeling to predict the hydromechanical behavior of the weakly consolidated rock is presented. This approach allows a satisfying prediction of the permeability evolution with stresses, using an exponential function of an effective strain. Reversely for DS, the link between strain and permeability is not obvious as permeability reduction is pronounced at early stage of loading. To understand these permeability evolutions, a damage analysis has been performed using core analysis measurements and multi-scale imaging.

O modelo cam-clay modificado foi aplicado aos resultados experimentais obtidos para um solo arenoso típico da cidade de São Carlos. Os ensaios de compressão triaxial foram conduzidos em equipamento moderno, com instrumentação interna, segundo distintas trajetórias de carregamento. Verificou-se que os resultados obtidos em termos de modelagem foram satisfatórios, principalmente quando a tensão octaédrica (p\') foi diminuida durante os carregamentos. Nesse caso, tanto em termos de modelagem como de resultados experimentais, houve expansão de volume do solo. Com o aumento da tensão octaédrica, verificou-se a ocorrência de compressão volumétrica do solo. Observou-se que o modelo apresenta uma previsão de deformações axiais maiores do que as observadas experimentalmente nas trajetórias de -30, -50, 30, 40, 50, 60, 120 graus e no ensaio triaxial convencional com \'sigma\' 3 = 100 kPa. Além disso, determinou-se a superfície inicial de plastificação do solo utilizando-se dois critérios que tenderam a fornecer valores de tensão de cedência aproximadamente iguais, notando-se que a condição de fluxo associado não é obedecida. / The modified cam-clay model was used to model experimental results of a sandy soil from São Carlos - SP. Triaxial compression tests were performed using Bishop - Wesley cell with internal transducers to measure axial and radial strains. It was observed that the model fairly fitted experimental results, specially when medium effective stress (p\') is reduced during loading. In this case, both the model and the experimental results, showed volume increase. When (p\') increases the model and the tests showed a tendency to give volumetric compression, although the values were differents. The model yielded strains larger than that measured in the tests when the stress-paths were of -30, -50, 30, 40, 50, 60, 120 degrees and in axial compression test with 100 kPa of confining pressure. Besides that, initial yield surface of soil was calculated from test results using two different criteria which gave about the same yield stress and it is show that normality rule was not satisfied in this soil.

O modelo cam-clay modificado foi aplicado aos resultados experimentais obtidos para um solo arenoso típico da cidade de São Carlos. Os ensaios de compressão triaxial foram conduzidos em equipamento moderno, com instrumentação interna, segundo distintas trajetórias de carregamento. Verificou-se que os resultados obtidos em termos de modelagem foram satisfatórios, principalmente quando a tensão octaédrica (p\') foi diminuida durante os carregamentos. Nesse caso, tanto em termos de modelagem como de resultados experimentais, houve expansão de volume do solo. Com o aumento da tensão octaédrica, verificou-se a ocorrência de compressão volumétrica do solo. Observou-se que o modelo apresenta uma previsão de deformações axiais maiores do que as observadas experimentalmente nas trajetórias de -30, -50, 30, 40, 50, 60, 120 graus e no ensaio triaxial convencional com \'sigma\' 3 = 100 kPa. Além disso, determinou-se a superfície inicial de plastificação do solo utilizando-se dois critérios que tenderam a fornecer valores de tensão de cedência aproximadamente iguais, notando-se que a condição de fluxo associado não é obedecida. / The modified cam-clay model was used to model experimental results of a sandy soil from São Carlos - SP. Triaxial compression tests were performed using Bishop - Wesley cell with internal transducers to measure axial and radial strains. It was observed that the model fairly fitted experimental results, specially when medium effective stress (p\') is reduced during loading. In this case, both the model and the experimental results, showed volume increase. When (p\') increases the model and the tests showed a tendency to give volumetric compression, although the values were differents. The model yielded strains larger than that measured in the tests when the stress-paths were of -30, -50, 30, 40, 50, 60, 120 degrees and in axial compression test with 100 kPa of confining pressure. Besides that, initial yield surface of soil was calculated from test results using two different criteria which gave about the same yield stress and it is show that normality rule was not satisfied in this soil.