Erika Angerer

Fractured reservoir modeling from seismic to simulator: A reality?

Characterizing carbonate reservoirs has traditionally been a difficult task when the reservoir performance is strongly influenced by an often poorly constrained fracture system. Studies have shown that the geometry of the fracture network is a dominant control of fractured reservoir performance. Direct fracture data only exist at the well and so a long running quest has been the accurate description of the geometry and impact of fractures in the interwell reservoir zone. A range of different approaches has been used to constrain the expected distribution of subseismic fractures in the reservoir, including such techniques as geomechanical modeling, palinplastic reconstruction, and curvature analysis. However, these techniques are all attempting to predict the distribution of fracture based upon indirect observations or assumptions. What is ideally required is an approach that can directly sample the whole reservoir in order to determine the properties of the fracture network.

On the Use of Geostatistical Filtering Techniques In Seismic Processing

When redundancy of seismic data exists factorial cokriging enables the estimation of (1) a common part, based on the common spatial behavior, and (2) the differences relative to the common part of the input data. Coléou (2002) first introduced the automatic implementation of factorial co-kriging (AFACK) as a filtering technique for the time-lapse (4D) processing sequence. It was specifically designed to optimize the critical time-lapse information such as the repeatability and the 4D seismic signature. However, over the last two years we have been developing new applications of this technique in very different processing environments. For example, applications providing data reduction, such as stacking or AVO and EI analysis, have a direct interest in the common part of consecutive offset cubes. Furthermore, we have successfully applied AFACK to more specialized problems such as the merging of OBC and streamer data or the decomposition of wide-azimuth data for fracture characterization.

Constraining Models of Fractured Reservoirs Using Seismic Anisotropy Maps, For Improved Reservoir Performance And Prediction

Traditionally, reservoir fracture models are matched to only primary fracture data at well locations, and thus suffer from increasing uncertainty away from wells. In this work the discrete fracture network (DFN) approach is used to show how sub-seismic fracture realizations can also be constrained to seismic data. This methodology provides a route by which the uncertainty in the distribution of reservoir permeability and storage can be significantly reduced, and helps make forward predictions from the simulator more accurate. It also provides a natural way of extracting value from seismic anisotropy measurements and using them to quantitatively control the reservoir model for field development and management.

Wide-azimuth Processing For Azimuthal Anisotropy Analysis

Azimuth-friendly data processing is mandatory in a reliable fracture characterisation workflow. Here, the effects of standard processing methodologies in the presence of azimuthal anisotropy are investigated on synthetic and real data examples. Common techniques for signal processing, statics calculations, imaging and velocity analysis are adapted for wide-azimuth, wide-offset P-wave data. We present a general processing sequence that preserves azimuthal anisotropy. Further, we compare anisotropy parameters derived from impedance inversion and azimuthal AVO on a Middle East data example. Introduction Observable effects of azimuthal anisotropy on seismic data are of second order compared to the geological background. A careful preservation of azimuthal amplitude and travel time variations is therefore crucial. One of the main questions in processing wide-azimuth data in the presence of azimuthal anisotropy is when to process the entire volume continuously as a single data set and when to split and process the data in azimuth-limited sectors. We devised a testing sequence using synthetic, anisotropic, wide-azimuth data. Processing steps are applied to the data with and without azimuth sectoring and the preservation of anisotropy is assessed. Data processing sequence Signal processing Of specific interest are transform-based signal processing techniques like 2D and 3D FK, and FX noise suppression methods. Figure 1 shows the comparison of the resulting amplitudes before and after application of FX deconvolution to the synthetic data. The solid lines are the exact amplitudes calculated using Ruger’s elliptic approximations of the anisotropic reflection coefficient (1998). The magnitude of the azimuthal variation increases with increasing incidence angle. Figure 1a shows the best fitting ellipses to the amplitudes after adding noise (S/N=0.5) and applying FX deconvolution in azimuth sectors. Figure 1b shows the resulting amplitudes after FX deconvolution treating the data as a continuous volume and thereby mixing different azimuths. It is evident that the azimuthal content of the data is preserved when the data are filtered in azimuth sectors. There is a general shift in the ellipses that is due to the added noise. In Figure 1b amplitudes are not correctly preserved, the azimuthal variation is smeared out and therefore anisotropy is not preserved. Using the same approach we also tested 2D FK filtering and τ-p transform. The results are similar to the ones shown in Figure 1. In order to preserve azimuthal variations, transform-based signal processing algorithms need to be applied to data split or sorted into azimuth sectors. A 3D FK cone filter preserves azimuthal anisotropic variations, and for this the data are processed continuously. Moreover, surface-consistent processing steps are applied to the data set continuously since source and receiver related near-surface effects are azimuth