Les bandes de déformation sont des structures géologiques se mettant en place dans les matériels granulaires présentant une forte porosité (>15%). Ces structures peuvent être compactantes ou dilatantes et peuvent présenter une composante cisaillante. À l’échelle microscopique, il est possible d’observer une réorientation des grains, une compaction ou un cisaillement intense peut entrainer la fracturation de ces derniers (cataclase) pour former une fine zone déformée modifiant ainsi la porosité et la perméabilité de la roche. Ces bandes ont un impact sur la circulation des fluides en formant des barrières ou des drains dans le réservoir. La formation de ces structures est étroitement liée à la tectonique et aux paramètres sédimentologiques du matériel hôte. Comprendre et pouvoir prédire le mode de mise en place, les orientations et la distribution de ces bandes est l’objectif principal de cette thèse.Dans cette étude nous avons analysé un site de bandes de déformation observé dans le bassin de Tremp, au sein de la formation d’Aren, dans la zone Central Sud-Pyrénéenne. Nous avons pu définir la nature de ces structures grâce à des analyses macro- et micro-structurales couplées à une étude d’anisotropie magnétique permettant de déduire la direction de raccourcissement à l’origine de la mise en place de ces bandes. Deux principaux types de bandes cataclastiques sont alors mis en évidence: (1) des bandes de compaction pures, perpendiculaires au raccourcissement et (2) des bandes de compaction à composante cisaillante, obliques à cette même direction de raccourcissement.En comparaison avec le calendrier tectonique de la région et des données d’enfouissement relatif dans le temps de la formation étudiée. Nous avons fait l’hypothèse que ces deux familles de bandes sont apparues à faible enfouissement (< 1 km de profondeur), soit peu de temps après le dépôt et associées à la croissance du pli du Sant Corneli-Boixols. De telles structures ne sont pas communes pour un enfouissement superficiel et le faciès calcarénitique est mis en cause pour expliquer leurs apparitions.La mise en place de simulations analytiques basées sur des résultats d’expérimentations géomécaniques ont ensuite permis de contraindre le régime tectonique, l’orientation et les valeurs de ces contraintes nécessaires à la formation de ces deux types de bandes par rapport à l’enfouissement et la croissance du chevauchement. Les valeurs de contraintes attendues sont ici très faibles dans le cas d’une déformation très précoce.Enfin, nous avons testé ces observations et nos hypothèses à des modélisations numériques dans lesquelles nous avons analysé l’impact de la croissance d’un chevauchement et d’un pli de propagation de rampe. La distribution des contraintes et les potentielles bandes de déformation mises en place au sein d’un réservoir poreux situé en avant de ce pli ont été étudiées. Nous avons alors montré que nos hypothèses d’apparition superficielles de bandes de déformation étaient dépendantes de la position des enveloppes de ruptures (elles-mêmes dépendantes de la lithologie de la roche). Pour expliquer la mise en place des bandes étudiée dans ce mémoire, une résistance mécanique très faible de la roche est nécessaire pour permettre de former des bandes à moins d’un kilomètre de profondeur. / Deformation bands are geological structures that occur in porous and granular material presenting a high porosity (>15%). These structures can be identified as compactive or dilatant, a shear component is also often observed. At the microscopic scale, it is possible to observe a grain rearrangement and an intense compaction and or shearing can lead to grain crushing (known as cataclasis), to form a thin deform zone that will modify the porosity and permeability of the rock. Deformation bands have a non-negligible impact on fluid flow, creating a barrier or a drain in the potential reservoir. The formation of such structures is mainly linked to the tectonic activity but also to the facies and other sedimentological parameters of the host rock. The understanding and the prediction of the occurrence and distribution of the bands is the main objective of this thesis.In this study we analyzed a deformation band site found in the Tremp basin, in the Aren formation localized in the South Central Pyrenean Zone. We defined the nature of these structures with macro and microstructural analysis and by adding a study of the magnetic anisotropy to constrain the shortening direction responsible to the band formation. We evidence two major types of bands showing different orientations and behavior: (1) Pure compaction bands (PCB), perpendicular to the shortening and (2) Shear enhanced compaction bands (SECB), oblique to the same shortening.In comparison with tectonic schedule in the studied area and time vs. burial data of the formation, we deducted that both types of bands took place at a shallow burial (<1km depth), which means short times after deposition. This localized deformation, showing mainly cataclasis, is associated to the growth of the Sant Corneli-Boixols fold and thrust belt. Such structures are not common at a shallow depth and we propose that the calcarenite facies of the host rock is the key factor to explain the band occurrence.Thereafter, we made analytical simulations based on geomechanical experimentations results that allowed us to constrain the stress state and orientations needed to create these structure and to determine the timing of formation compared to the burial of the layers during the growth of the Boixols thrust. The stresses magnitudes are expected to be really low in the case of an early deformation.Finally, we tested and compared our observations and hypothesis to numerical modeling where we analyzed the impact of the growth of a fold and thrust belt on the stress state and orientations and the analysis of potential deformation bands occurrence. The stress distribution and the potential occurrence of deformation bands in a porous reservoir presenting different characteristics and located in front of this fold were studied.With the modelizations results, we exposed that our hypothesis of shallow deformation bands are dependent from the position of failure envelopes (that are dependent on the rock lithology). To explain the band formation we studied in this thesis, a weak mechanical strength of the host rock is needed to form deformation bands at less than a depth of one kilometer. The pure compaction bands are associated to a potentially early layer-parallel shortening (LPS).

In this dissertation, 3D surface seismic data was integrated with the well logs to be able to define the properties in every location for the reservoir under investigation. To accomplish this task, geostatistical and artificial neural networks (ANN) techniques were employed. First, missing log sets in the study area were estimated using common empirical relationships and ANN. Empirical estimations showed linear dependent results that cannot be generalized. On the other hand, ANNs predicted missing logs with an very high accuracy. Sonic logs were predicted using resistivity logs with 90% correlation coefficient. Second, acoustic impedance property was predicted in the study area. AI estimation first performed using sonic log with GRNN and 88% CC was obtained. AI estimation was repeated using sonic and resistivity logs and the result were improved to 94% CC. In the final part of the study, SGS technique was used with collocated cokriging techniques to estimate NPHI property. Results were varying due to nature of the algorithm. Then, GRNN and RNN algorithms were applied to predict NPHI property. Using optimized GRNN network parameters, NPHI was estimated with high accuracy. Results of the study were showed that ANN provides a powerful solution for reservoir parameter prediction in the study area with its flexibility to find out nonlinear relationships from the existing available data.

In South Central Cut Bank Sand Unit (SCCBSU) of Cut Bank field, primary production and waterflood projects have resulted in recovery of only 29 % of the original oil in place from heterogeneous, fluvial sandstone deposits. Using highresolution sequence stratigraphy and geostatistical analysis, I developed a geologic model that may improve the ultimate recovery of oil from this field. In this study, I assessed sequence stratigraphic concepts for continental settings and extended the techniques to analyze low-accommodation alluvial systems of the Cut Bank and Sunburst members of the lower Kootenai formation (Cretaceous) in Cut Bank field. Identification and delineation of five sequences and their bounding surfaces led to a better understanding of the reservoir distribution and variability. Recognition of stacking patterns allowed for the prediction of reservoir rock quality. Within each systems tract, the best quality reservoir rocks are strongly concentrated in the lowstand systems tract. Erosional events associated with falling baselevel resulted in stacked, communicated (multistory) reservoirs. The lowermost Cut Bank sandstone has the highest reservoir quality and is a braided stream parasequence. Average net-to-gross ratio value (0.6) is greater than in other reservoir intervals. Little additional stratigraphically untapped oil is expected in the lowermost Cut Bank sandstone. Over most of the SCCBSU, the Sunburst and the upper Cut Bank strata are valley-fill complexes with interfluves that may laterally compartmentalize reservoir sands. Basal Sunburst sand (Sunburst 1, average net-to-gross ratio ~0.3) has better reservoir quality than other Sunburst or upper Cut Bank sands, but its reservoir quality is significantly less than that of lower Cut Bank sand. Geostatistical analysis provided equiprobable representations of the heterogeneity of reservoirs. Simulated reservoir geometries resulted in an improved description of reservoir distribution and connectivity, as well as occurrences of flow barriers. The models resulting from this study can be used to improve reservoir management and well placement and to predict reservoir performance in Cut Bank field. The technical approaches and tools from this study can be used to improve descriptions of other oil and gas reservoirs in similar depositional systems.

The quantitative estimation of reservoir properties directly from seismic data is a major goal of reservoir characterization. Integrated reservoir characterization makes use of different varieties of well and seismic data to construct detailed spatial estimates of petrophysical and fluid reservoir properties. The advantage of data integration is the generation of consistent and accurate reservoir models that can be used for reservoir optimization, management and development. This is particularly valuable in mature field settings where hydrocarbons are known to exist but their exact location, pay, lateral variations and other properties are poorly defined. Recent approaches of reservoir characterization make use of individual seismic attributes to estimate inter-well reservoir properties. However, these attributes share a considerable amount of information among them and can lead to spurious correlations. An alternative approach is to evaluate reservoir properties using multiple seismic attributes. This study reports the results of an investigation of the use of multivariate seismic attributes to predict lateral reservoir properties of gross thickness, net thickness, gross effective porosity, net-to-gross ratio and net reservoir porosity thickness product. This approach uses principal component analysis and principal factor analysis to transform eighteen relatively correlated original seismic attributes into a set of mutually orthogonal or independent PC??s and PF??s which are designated as multivariate seismic attributes. Data from the N-sand interval of Vermilion Block 50 field, Gulf of Mexico, was used in this study. Multivariate analyses produced eighteen PC??s and three PF??s grid maps. A collocated cokriging geostaistical technique was used to estimate the spatial distribution of reservoir properties of eighteen wells penetrating the N-sand interval. Reservoir property maps generated by using multivariate seismic attributes yield highly accurate predictions of reservoir properties when compared to predictions produced with original individual seismic attributes. To the contrary of the original seismic attribute results, predicted reservoir properties of the multivariate seismic attributes honor the lateral geological heterogeneities imbedded within seismic data and strongly maintain the proposed geological model of the N-sand interval. Results suggest that multivariate seismic attribute technique can be used to predict various reservoir properties and can be applied to a wide variety of geological and geophysical settings.

The quantitative estimation of reservoir properties directly from seismic data is a major goal of reservoir characterization. Integrated reservoir characterization makes use of different varieties of well and seismic data to construct detailed spatial estimates of petrophysical and fluid reservoir properties. The advantage of data integration is the generation of consistent and accurate reservoir models that can be used for reservoir optimization, management and development. This is particularly valuable in mature field settings where hydrocarbons are known to exist but their exact location, pay, lateral variations and other properties are poorly defined. Recent approaches of reservoir characterization make use of individual seismic attributes to estimate inter-well reservoir properties. However, these attributes share a considerable amount of information among them and can lead to spurious correlations. An alternative approach is to evaluate reservoir properties using multiple seismic attributes. This study reports the results of an investigation of the use of multivariate seismic attributes to predict lateral reservoir properties of gross thickness, net thickness, gross effective porosity, net-to-gross ratio and net reservoir porosity thickness product. This approach uses principal component analysis and principal factor analysis to transform eighteen relatively correlated original seismic attributes into a set of mutually orthogonal or independent PC??s and PF??s which are designated as multivariate seismic attributes. Data from the N-sand interval of Vermilion Block 50 field, Gulf of Mexico, was used in this study. Multivariate analyses produced eighteen PC??s and three PF??s grid maps. A collocated cokriging geostaistical technique was used to estimate the spatial distribution of reservoir properties of eighteen wells penetrating the N-sand interval. Reservoir property maps generated by using multivariate seismic attributes yield highly accurate predictions of reservoir properties when compared to predictions produced with original individual seismic attributes. To the contrary of the original seismic attribute results, predicted reservoir properties of the multivariate seismic attributes honor the lateral geological heterogeneities imbedded within seismic data and strongly maintain the proposed geological model of the N-sand interval. Results suggest that multivariate seismic attribute technique can be used to predict various reservoir properties and can be applied to a wide variety of geological and geophysical settings.