Pierre Thore

Structural uncertainties: Determination, management, and applications

Structural uncertainties have a direct impact in exploration, development, and production, and in drilling decisions. In this paper, we present an approach for determining and handling structural uncertainties. We first examine the magnitude of the different sources of uncertainty, and explain how to estimate their direction and correlation length. This task requires a huge geophysical input. This information is then used in a general scheme to generate multiple realizations of the structural model consistent with structural uncertainties. The technique is based on geostatistical concepts. Finally, we illustrate the application of this scheme in examples relevant for exploration, development and production, and drilling.The structural model is described as a set of horizons represented by triangulated surfaces cut by faults. The relationships between horizons and faults are expressed as a set of constraints. On a horizon, each source of uncertainty (typically migration, picking, and time‐to‐depth conversi...

Time-lapse seismic imaging using regularized full-waveform inversion with a prior model: which strategy?

Full-waveform inversion is an appealing technique for time-lapse imaging, especially when prior model information is included into the inversion workflow. Once the baseline reconstruction is achieved, several strategies can be used to assess the physical parameter changes, such as parallel difference (two separate inversions of baseline and monitor data sets), sequential difference (inversion of the monitor data set starting from the recovered baseline model) and double-difference (inversion of the difference data starting from the recovered baseline model) strategies. Using syntheticMarmousi data sets, we investigate which strategy should be adopted to obtain more robust and more accurate time-lapse velocity changes in noise-free and noisy environments. This synthetic application demonstrates that the double-difference strategy provides the more robust time-lapse result. In addition, we propose a target-oriented time-lapse imaging using regularized full-waveform inversion including a prior model and model weighting, if the prior information exists on the location of expected variations. This scheme applies strong prior model constraints outside of the expected areas of timelapse changes and relatively less prior constraints in the time-lapse target zones. In application of this process to the Marmousi model data set, the local resolution analysis performed with spike tests shows that the target-oriented inversion prevents the occurrence of artefacts outside the target areas, which could contaminate and compromise the reconstruction of the effective time-lapse changes, especially when using the sequential difference strategy. In a strongly noisy case, the target-oriented prior model weighting ensures the same behaviour for both time-lapse strategies, the double-difference and the sequential difference strategies and leads to a more robust reconstruction of the weak time-lapse changes. The double-difference strategy can deliver more accurate time-lapse variation since it can focus to invert the difference data. However, the double-difference strategy requires a preprocessing step on data sets such as time-lapse binning to have a similar source/receiver location between two surveys, while the sequential difference needs less this requirement. If we have prior information about the area of changes, the target-oriented sequential difference strategy can be an alternative and can provide the same robust result as the doubledifference strategy.

Modelling of stochastic faults and fault networks in a structural uncertainty study

Modelling faults from seismic data for a 3D depth model is a difficult task because of the multiple sources of uncertainty. The uncertainty may be attributed to migration velocities, picking of faults and organization of the fault network in 3D. Faults are generally not migrated from time to depth domain like horizons are, but modelled in the depth domain from the depth migrated horizons. For this reason, a new data structure has been designed that is targeted for fault modelling. Taking uncertainties into account, this structure allows for rapid modelling of faults from depth migrated horizons. The input data and the parameterization of the new data structure will be described. Following this, a way to incorporate uncertainties during the interpretation process is proposed and a description of different stochastic methods used to compute new shapes and locations inside a given uncertainty volume will be made. Finally, the method and the results obtained will be described while studying uncertainties on more complex fault networks. The influence of fault uncertainties on the reservoir volumetric estimates will be shown as one possible result of the simulation process.

Sensitivity Analysis of Time-lapse Images Obtained By Differential Waveform Inversion With Respect to Reference Model

SUMMARYFor monitoring purposes, one of the promising techniques ded-icated to assess physical properties changes in target regionsis the differential waveform inversion, both in the acoustic andelastic cases. A central question of this technique regards thechoice of the reference model. One solution could be the useof the reconstructed baseline image provided through the stan-dard Full Waveform Inversion (FWI) procedure of initial data.However, how the accuracy of the baseline reconstructed imagewill affect the precision of further time-lapse images is of cru-cial importance. Here, we present a sensitivity analysis of time-lapse images obtained from differential inversion, with respectto various reference models. Density, P- and S-wave velocitychanges could be converted into fluid property changes thanksto an empirical downscaling relationship. For accurate estima-tion of fluid parameter changes, the construction of highly re-solved time-lapse images presenting acceptable errors is a keyissue for the downscaling procedure. We illustrate on a specificsynthetic example that the sensitivity analysis over the refer-ence model variation provides linear convergence towards thetime-lapse image obtained when using the exact baseline. Anaccurate baseline reconstruction is essential and could benefitfrom other data collected for monitoring purposes.INTRODUCTIONFWI is a data fitting procedure aiming to develop high resolu-tion quantitative images of the subsurface, through the extrac-tion of the full information content of the seismic data (Taran-tola, 1984). Beside the exploration application, the FWI methodcan be also used for monitoring applications, such as oil and gasreservoirs, steam injection, CO

Application of a grid-consistent inversion to a 4D reservoir model

Extraction of the 4D signal from time-lapse seismic is an efficient way to evaluate the lateral evolution of changes in fluid pressure and saturation during hydrocarbon production. This 4D seismic signal can be used in two ways (Castro et al., 2009): qualitatively (from a static perspective), thus evaluating whether a compartment has been drained or not, or if a fault is permeable or not; or quantitatively (from a dynamic perspective) by trying to exactly match changes in dynamic properties predicted from fluid-flow simulation with 4D seismic signal itself. In both approaches, the main difficulty is to link the seismic information with a description of the reservoir grid, especially with regard to scaling problems. Relative to the simulation model, the 4D seismic signal is obtained with a higher lateral resolution (12.5 or 25 m) and a poorer vertical resolution (10–20 m). Indeed, the reservoir grid consists of a set of layers with a very high vertical resolution (1–5 m) and limited lateral resolution (50–...

Towards a better seismic to well tie in complex media

Summary The conventional approach for seismic to well tying has proven to be of high efficiency in most cases when the bas ic ingredients, as documented by White (1998), are met: high quality seismic data and well logs, and near-vertical wells. Structural complexity can considerably affect the result of the seismic to well calibration process. Highly dipping strata, highly deviated wells, rapid lateral velocity variations, or anisotropy lie in many cases behind a poor seismic to well tie. In this paper, we study the various parameters that stand in the way of a good tie in complex media. In a first stage, we used a highly realistic modellingimaging ch ain to better understand “where, how and why” the synthetics computed from the well differ from the seismic data. Starting from the simplest case and complexifying the structural model in various ways, we analyse where and when the traditional approach fails. Conclusions will be drawn from this study and various examples wil l be presented.

Accuracy and limitations in seismic modeling of reservoir

We use a real West African deep offshore reservoir for checking the accuracy of common practice in seismic modeling. In a first stage we examine the impact of transforming the reservoir/geological grid into a regular grid suitable for seismic modeling. This operation is not trivial, it implies upscaling and downscaling aspects and depending on the algorithms used, it conducts to variations which have a dramatic impact on synthetics. In a second stage we examine the limitations due to using convolution plus Zoeppritz equation for modeling. We compare the results obtained with this technique with those obtained with Finite Differences modeling: the synthetics show important differences. From an operational point of view this study demonstrates that we should question the reliability of our inversion results.

Tying seismic to well data using structural uncertainties

In order to tie final depth migrated interpretations to well locations, one method used is often to compute a vertical stretching map and apply it to the depth image. The authors want to propose here an alternative method for this tying, using principal uncertainty directions to decide how to move each individual surface element so that the global interpretations fits the well data. These uncertainty directions are an approximation of the displacement vector of a surface element for a given velocity error. They are used for both associating each well location to a surface location, and for estimating a velocity error coefficient to be applied on the interpretations. Extensive use of the G{bigcirc}CAD software is made throughout the process. It is used for the preliminary steps of building the depth surfaces, and allows the attachment of user-defined properties, such as the uncertainty vector coordinates and the velocity error coefficient, to be stored on the surface locations as well. It also provides the basic tools needed for 3D proximity checking, as well as a powerful built-in interpolation method.

Modeling of seismic wave propagation in presence of topography

In this work we studied and developed an efficient method to handle the wave simulation in presence of topography based on curvilinear finite difference. Foremost, we derived the modified equations and the free surface condition on the continuous problem. We used afterwards optimized stencils and optimized selective filters adapted from aeroacoustics. The use of conventional grid allowed us to directly extend the non-centered stencils at the boundaries developed by Berland et al. (2007) to our problem in curvilinear coordinates.

Well-to-seismic tie method in complex imaging areas Examples in the deep offshore subsalt Angola domain

Recently, imaging capabilities have allowed a major step forward in subsalt exploration. In particular, new PSDM algorithms have appeared on the market that account for complex wave propagation through salt canopies to produce better images. These improvements have allowed the drilling of numerous subsalt exploration wells, but at the time of appraisal and development, questions arise on the reliability of these seismic cubes to accurately predict geometries. This issue is crucial in the offshore deepwater blocks of Angola, where reflectors often have high dips and subsalt seismic visibility is not optimal. Furthermore, velocity models used for subsalt imaging are seldom as accurate as one could want, and significant discrepancies can often be observed between seismic and well depth of geological markers. These discrepancies can have several origins, such as neglecting to account for anisotropy, as discussed by Shumaker et al. (2007). The lack of accuracy of velocity models can also be due to the limitati...