Production from a hydrocarbon reservoir is typically supported by water or carbon dioxide (CO2) injection. CO2 injection into hydrocarbon reservoirs is also a promising solution for reducing environmental hazards from the release of green house gases into the earth’s atmosphere. Numerical simulators are used for designing and predicting the complex behavior of systems under such scenarios. Two key steps in such studies are forward modeling for performance prediction based on simulation studies using reservoir models and inverse modeling for updating reservoir models using the data collected from field. The viability of time-lapse seismic monitoring using an integrated modeling of fluid flow, including chemical reactions, and seismic response is examined. A comprehensive simulation of the gas injection process accounting for the phase behavior of CO2-reservoir fluids, the associated precipitation/dissolution reactions, and the accompanying changes in porosity and permeability is performed. The simulation results are then used to model the changes in seismic response with time. The general observation is that gas injection decreases bulk density and wave velocity of the host rock system. Another key topic covered in this work is the data assimilation study for hydrocarbon reservoirs using Ensemble Kalman Filter (EnKF). Some critical issues related to EnKF based history matching are explored, primarily for a large field with substantial production history. A novel and efficient approach based on spectral clustering to select ‘optimal’ initial ensemble members is proposed. Also, well-specific black-oil or compositional streamline trajectories are used for covariance localization. Approach is applied to the Weyburn field, a large carbonate reservoir in Canada. The approach for optimal member selection is found to be effective in reducing the ensemble size which was critical for this large-scale field application. Streamline-based covariance localization is shown to play a very important role by removing spurious covariances between any well and far-off cell permeabilities. Finally, time-lapse seismic study is done for the Weyburn field. Sensitivity of various bulk seismic parameters viz velocity and impedance is calculated with respect to different simulation parameters. Results show large correlation between porosity and seismic parameters. Bulk seismic parameters are sensitive to net overburden pressure at its low values. Time-lapse changes in pore-pressure lead to changes in bulk parameters like velocity and impedance.

Sequestration has been suggested as a solution for resolving the problem of increasing greenhouse gas emissions. CO2 is the major greenhouse gas which results from using fossil fuels for domestic and industrial purposes. Different geological targets have been suggested as reservoirs for CO2 sequestration with saline aquifers being the focus of this research. Monitoring and verification of injected CO2 into the ground is an essential part of CO2 sequestration because there is a strong requirement to understand and correctly manage the CO2 flow and movement within the reservoir over time. This includes a need to understand mobile CO2 in its all phases (gas, liquid, supercritical and dissolved in formation water). It is now well recognised that monitoring injected liquids in the sub-surface can be done remotely using surface seismic monitoring techniques. Seismic waves are sensitive to the contrast in the physical properties of formation water and CO2. As a gas, the migration path of CO2 has been shown to be easily imaged but such images provide only a qualitative rather than a quantitative solution, which is inadequate to remotely verify storage volumetrics. The complexity of saline aquifer reservoirs containing the different phases of CO2 (a function of reservoir pressure, temperature, and chemical composition and the state of phase of injected CO2) requires a good knowledge base of how the seismic response changes to such changes in CO2 phase and reservoir heterogeneities for verification purposes. / In this research, transmission ultrasonic seismic experiments were performed under controlled pressure, temperature and CO2 dissolution conditions in water. Different forms of simulated rock matrix were used to understand how seismic attributes changed with changing sequestration conditions. Data analysis showed that the commonly used approach of seismic velocity analysis is not particularly sensitive to dissolved CO2 whereas seismic amplitude was very sensitive to dissolved CO2 content and is the seismic attribute of choice for the future quantification of CO2. The density increase in formation water brine as a result of CO2 mixture was found to be directly related to transmission amplitude and provides the potential for prediction and thus, remote quantification. Also, there was confirmation during the transmission experiments that seismic amplitude changes markedly when CO2 changes phase from its dissolved form into a gas, as a result of significant attenuation by CO2 bubbles. Analysis showed that the dominant and centre frequency of the spectra also responded to CO2 phase when it changed from dissolved to its free gas form. However, these attributes appear to be of use in a qualitative manner rather than quantitative. The CO2 pre-bubble phase was studied in an attempt to obtain a basic knowledge of the effect on seismic amplitude variation for quantifying dissolved gas amounts with some success. This knowledge has an application in Gas-to-Oil-Ratio mapping in depleting oil fields and can assist the future management of production from fields which are at the stage of near-bubble point due to pressure depletion. / The results of this research have an application in time-lapse seismic monitoring and operational management of greenhouse gas sequestration operations. In particular, the VSP and cross-well seismic methods are immediate beneficiaries of this research, with further work required for application to 3-D reflectivity methods in time-lapse surface seismic monitoring.

Time-lapse seismic allows oil/gas reservoir monitoring during production, highlighting compaction and water movement. Time-lapse modeling, using a stress-dependent rock physics model, helps determine the need and frequency of expensive repeat seismic acquisition. We simulate a Gulf of Mexico gas reservoir time-lapse response for depletion and water flooding using uncertainty ranges in water saturation, porosity, stress-induced velocity changes, and pore compressibility. An analysis is conducted to see if a water-swept region could have been predicted. Findings show the swept and un-swept monitor cases amplitude differences range from 6% to 15%, which is higher than the actual monitor seismic noise level. Thus, it is unlikely these cases could be differentiated. However, the modeled amplitude changes from base to monitor cases do not match measured amplitude changes. This suggests the rock property model requires pressure-variance improvement and/or the changes in seismic amplitudes are associated with pressure/porosity, thickness, or saturation cases not modeled.

Seismic full waveform inversion (waveform tomography) is a method to reconstruct the underground velocity field in high resolution using seismic data. The method was first introduced during the 1980’s and became computationally feasible during the late 1990’s when the method was implemented in the frequency domain. This work presents three case studies and one synthetic benchmark of full waveform inversion applications. Two of the case studies are focused on time-lapse cross-well and 2D reflection seismic data sets acquired at the Ketzin CO2 geological storage site. These studies are parts of the CO2SINK and CO2MAN projects. The results show that waveform tomography is more effective than traveltime tomography for the CO2 injection monitoring at the Ketzin site for the cross-well geometry. For the surface data sets we find it is difficult to recover the true value of the velocity anomaly due to the injection using the waveform inversion method, but it is possible to qualitatively locate the distribution of the injected CO2. The results agree well with expectations based upon conventional 2D CDP processing methods and more extensive 3D CDP processing methods in the area. A further investigation was done to study the feasibility and efficiency of seismic full waveform inversion for time-lapse monitoring of onshore CO2 geological storage sites using a reflection seismic geometry with synthetic data sets. The results show that waveform inversion may be a good complement to standard CDP processing when monitoring CO2 injection. The choice of method and strategy for waveform inversion is quite dependent on the goals of the time-lapse monitoring of the CO2 injection. The last case study is an application of the full waveform inversion method to two crooked profiles at the Forsmark site in eastern central Sweden. The main goal of this study was to help determine if the observed reflections are mainly due to fluid filled fracture zones or mafic sills. One main difficulty here is that the profiles have a crooked line geometry which corresponds to 3D seismic geometry, but a 2D based inversion method is being used. This is partly handled by a 3D to 2D coordinate projection method from traveltime inversion. The results show that these reflections are primarily due to zones of lower velocity, consistent with them being generated at water filled fracture zones.

Time-lapse seismic monitoring repeats 3D seismic imaging over a reservoir to map fluid movements in a reservoir. During hydrocarbon production, the fluid saturation, pressure, and temperature of a reservoir change, thereby altering the acoustic properties of the reservoir. Time-lapse seismic analysis can illuminate these dynamic changes of reservoir properties, and therefore has strong potential for improving reservoir management. However, the response of a reservoir depends on many parameters and can be diffcult to understand and predict. Numerical modeling results integrating streamline fluid flow simulation, rock physics, and ray-Born seismic modeling address some of these problems. Calculations show that the sensitivity of amplitude changes to porosity depend on the type of sediment comprising the reservoir. For consolidated rock, high-porosity models show larger amplitude changes than low porosity models. However, in an unconsolidated formation, there is less consistent correlation between amplitude and porosity. The rapid time-lapse modeling schemes also allow statistical analysis of the uncertainty in seismic response associated with poorly known values of reservoir parameters such as permeability and dry bulk modulus. Results show that for permeability, the maximum uncertainties in time-lapse seismic signals occur at the water front, where saturation is most variable. For the dry bulk-modulus, the uncertainty is greatest near the injection well, where the maximum saturation changes occur. Time-lapse seismic methods can also be applied to monitor CO2 sequestration. Simulations show that since the acoustic properties of CO2 are very different from those of hydrocarbons and water, it is possible to image CO2 saturation using seismic monitoring. Furthermore, amplitude changes after supercritical fluid CO2 injection are larger than liquid CO2 injection. Two seismic surveys over Teal South Field, Eugene Island, Gulf of Mexico, were acquired at different times, and the numerical models provide important insights to understand changes in the reservoir. 4D seismic differences after cross-equalization show that amplitude dimming occurs in the northeast and brightening occurs in the southwest part of the field. Our forward model, which integrates production data, petrophysicals, and seismic wave propagation simulation, shows that the amplitude dimming and brightening can be explained by pore pressure drops and gas invasion, respectively.

Time-lapse seismic is a modern technology for monitoring production-induced changes in and around a hydrocarbon reservoir. Time-lapse (4D) seismic may help locate undrained areas, monitor pore fluid changes and identify reservoir compartmentalization. Despite several successful 4D projects, there are still many challenges related to time-lapse technology. Perhaps the most important are to perform quantitative time-lapse and to model and interpret time-lapse effects in thin layers. The former requires one to quantify saturation and pressure effects on rock elastic parameters. The latter requires an understanding of the combined response of time-lapse effects in thin layers and overcoming seismic vertical resolution limitation. / This thesis presents an integrated study of saturation and pressure effects on elastic properties. Despite the fact that Gassmann fluid substitution is standard practice to predict time-lapse saturation effects, its validity in the field environment rests upon a number of assumptions. The validity of Gassmann equations, ultimately, can only be tested in real geological environments. In this thesis I developed a workflow to test Gassmann fluid substitution by comparing saturated P-wave moduli computed from dry core measurements with those obtained from sonic and density logs. The workflow has been tested on a turbidite reservoir from the Campos Basin, offshore Brazil. The results show good statistical agreement between the P-wave elastic moduli computed from cores using the Gassmann equations and the corresponding moduli computed from log data. This confirms that all the assumptions of the Gassmann theory are adequate within the measurement error and natural variability of elastic properties. These results provide further justification for using the Gassmann theory to interpret time-lapse effects in this sandstone reservoir and in similar geological formations. / Pressure effects on elastic properties are usually obtained by laboratory measurements, which can be affected by core damage. I investigated the magnitude of this effect on compressional-wave velocities by comparing laboratory experiments and log measurements. I used Gassmann fluid substitution to obtain low-frequency saturated velocities from dry core measurements taken at reservoir pressure, thus mitigating the dispersion effects. The analysis is performed for an unusual densely cored well from which 43 cores were extracted over a 45 m thick turbidite reservoir. These computed velocities show very good agreement with the sonic-log measurements. This is encouraging because it implies that core damages that may occur while bringing the core samples to the surface are small and do not adversely affect the measurement of elastic properties on these core samples. Should core damage have affected our measurements, we would have expected a systematic difference between properties measured in situ and on the recovered. This confirms that, for this particular region, the effect of core damage on ultrasonic measurements is less than the measurement error. Consequently, stress sensitivity of elastic properties as obtained from ultrasonic measurements are adequate for quantitative interpretation of time-lapse seismic data. / In some circumstances, stress sensitivity may not be obtained by ultrasonic measurements. Cores may be affected by damage, bias in the plugging process and scale effects and therefore may not be representative of the in situ properties. Consequently it is desirable to obtain this dependence from an alternative method. This other approach ideally should provide the pressure - velocity dependence from an intact rock. Few methods can sample the in situ rock. Seismic, for instance, provides in situ information, but lacks vertical resolution. Well logs, on the other hand, can provide high vertical resolution information, but usually are not available before and after production changes. I propose a method to assess the in situ pressure - velocity dependence using well data. I apply this method to a reservoir made up of sandstone. I used 23 wells drilled and logged in different stages of development of a hydrocarbon field providing rock and fluid properties at different pressures. For each well logged at a specific time, pore pressure, velocity and porosity, among other properties, are known. Pore pressure is accessed from a Repeat Formation Tester (RFT). As a field depletes and new wells are drilled and logged, similar data sets related to different stages of depletion are available. I present an approach expanding Furre et al. (2009) study incorporating porosity and obtaining a three dimensional relationship with velocity and pressure. The idea is to help to capture rock property variability. / Quantitative time-lapse studies require precise knowledge of the response of rocks sampled by a seismic wave. Small-scale vertical changes in rock properties, such as those resulting from centimetre scale depositional layering, are usually undetectable in both seismic and standard borehole logs (Murphy et al., 1984). I present a methodology to assess rock properties by using X-ray computed tomography (CT) images along with laboratory velocity measurements and borehole logs. This methodology is applied to rocks extracted from around 2.8 km depth from offshore Brazil. This improved understanding of physical property variations may help to correlate stratigraphy between wells and to calibrate pressure effects on velocities, for seismic time-lapse studies. / Small scale intra-reservoir shales have a very different response from sands to fluid injection and depletion, and thus may have a strong effect on the equivalent properties of a heterogeneous sandstone reservoir. Since shales have very low permeability, an increase of pore pressure in the sand will cause an increase of confining pressure in the intra-reservoir shale. I present a methodology to compute the combined seismic response for depletion and injection scenarios as a function of net to gross (NTG or sand – shale fraction). This approach is appropriate for modelling time-lapse effects of thin layers of sandstones and shales in repeated seismic surveys when there is no time for pressure in shale and sand to equilibrate. I apply the developed methodology to analyse the sand - shale combined response to typical shale and sandstone stress sensitivities for an oil field located in Campos Basin, Brazil. For a typical NTG of 0.6, there is a difference of approximately 35% in reflection coefficient during reservoir depletion from the expected value if these shales are neglected. Consequently, not considering the small shales intra-reservoir may mislead quantitative 4D studies. / The results obtained in this research are aimed to quantify pressure and saturation effects on elastic properties. New methodologies and workflows have been proposed and tested using real data from South America (Campos Basin) datasets. The results of this study are expected to guide future time-lapse studies in this region. Further investigations using the proposed methodologies are necessary to verify their applicability in other regions.

Seismic inversion is the leading method to map and quantify changes in time-lapse (4D) seismic datasets, with applications ranging from monitoring hydrocarbon-producing fields to geological CO2 storage. However, the process of inverting seismic data for reservoir properties is a notoriously ill-posed inverse problem due to the band-limited and noisy nature of seismic data. This comes with additional challenges for 4D applications, given the inaccuracies in the repeatability of time-lapse acquisition surveys. Consequently, adding prior information to the inversion process in the form of properly crafted regularization terms is essential to obtain geologically meaningful subsurface models and 4D effects. In this thesis, I propose a joint inversion-segmentation algorithm for 4D seismic inversion, which integrates total variation and segmentation priors as a way to counteract the missing frequencies and noise present in 4D seismic data. I validate the algorithm with synthetic and field seismic datasets and benchmark it against state-of-the-art 4D inversion techniques. The proposed algorithm shows three main advantages: 1. it produces high-resolution baseline and monitor acoustic impedance models, 2. by leveraging similarities between multiple seismic datasets, the proposed algorithm mitigates the non-repeatable noise and better highlights the real seismic time-lapse changes, and 3. it simultaneously provides a volumetric classification of the acoustic impedance 4D difference model based on user-defined classes, i.e., percentages of seismic time-lapse changes. Such advantages may enable more robust stratigraphic/structural and quantitative 4D seismic interpretation and provide more accurate inputs for dynamic reservoir simulations.

L’objectif de la thèse est de développer une méthodologie permettant d’obtenir une interprétation quantitative des données de sismique répétée (sismique 4D). Une étape essentielle consiste à évaluer la répétitivité des données, puis à filtrer les bruits indésirables, qui peuvent masquer ou détériorer la signature 4D. Une méthodologie basée sur des outils géostatistiques a été développée. Deux fenêtres temporelles des cubes sismiques étudiés sont choisies, l’une au-dessus du réservoir - où aucun effet 4D n’est attendu - et l’autre incluant le réservoir. Une analyse statistique et variographique conduite sur ces différentes fenêtres permet de proposer une décomposition des variogrammes en structures communes ou indépendantes, qui sont ensuite interprétées en termes de bruit, de signature géologique ou 4D. Les structures interprétées comme indépendantes de la géologie ou de la production sont ensuite filtrées à l’aide de la technique du krigeage factoriel proposée par Matheron en 1982. La méthodologie a été testée sur deux cas réels. Le premier concerne un champ d’huiles lourdes canadien, sur lequel trois campagnes sismiques ont été enregistrées, avant et pendant la production obtenue à l’aide d’injection de vapeur. Le bruit interprété comme tel sur les 3 campagnes a été filtré à l’aide la méthode décrite plus haut ; une analyse en termes de faciès sismiques non supervisée a ensuite été conduite sur les données brutes et filtrées afin d’évaluer l’intérêt du filtrage. Finalement, une interprétation des décalages en temps observés entre campagnes a été proposée. Le deuxième cas réel concerne un champ turbiditique profond dans l’offshore du Brésil, sur lequel deux campagnes sismiques 3D ont été acquises, avant et après le début de la production obtenue par injection d’eau. Le bruit présent sur les deux campagnes a été filtré à l’aide de la technique du krigeage factoriel, et la qualité de ce filtrage a pu être évaluée par comparaison avec d’autres techniques couramment utilisées / This thesis research aims at investigating seismic interpretation methodologies and techniques that will help on better characterizing time-lapse, or 4D, seismic signatures. These techniques and methodologies are used to evaluate the time-lapse repeatability and then to filter out undesirable artefacts that are non-related to the production, while enhancing the 4D signature. To achieve these goals, a methodology based on geostatistical tools, was developed. Typically, at least two time-interval windows are considered: one above and the other comprising the reservoir of interest. A statistical and variographic analysis, conducted on both windows and on all surveys, leads to an interpretation step where common or independent structures – in the variographic sense- can be pointed out. The structures interpreted as not related to the geology or to the production mechanism are filtered from the data by a multivariate factorial cokriging technique, based on the concept of Kriging Analysis developed by Matheron in 1982. Two real case time-lapse studies were used to test the methodology. The first case is a Canadian onshore heavy oil reservoir submitted to steam injection, where three different time-lapse surveys were shot to monitor the steam-chamber evolution. The noise present in the three surveys was first filtered using the technique described above; next, an unsupervised seismic facies analysis was conducted on both raw and filtered data in order to evaluate the filtering technique, and finally an interpretation, in terms of reservoir properties changes, of the time-shit observed between the campaigns was proposed. In the second case, the seismic data was acquired on a deepwater turbiditic oilfield from Brazil at two different times of reservoir life, before and after production and water injection. The two seismic surveys were filtered using the factorial kriging technique; the quality of the filtering was, in this case, evaluated by comparison with more common techniques