Mallick, Tanmay (Shell India Markets Private Limited) | Garg, Ashutosh (Shell India Markets Private Limited) | Choudhary, Manish (Shell India Markets Private Limited) | Nair, Saritha (Shell India Markets Private Limited) | Pal, Sabyasachi (Shell India Markets Private Limited) | Jana, Debadrita (Shell India Markets Private Limited) | Singh, Abhinav (Shell India Markets Private Limited) | Goudswaard, Jeroen (Shell India Markets Private Limited) | Faulkner, Andrew (Shell India Markets Private Limited) | Salakhetdinov, Ravil (Shell India Markets Private Limited)
A new seismic and quantitative reinterpretation was carried out for a brownfield in Western Desert, Egypt to improve depth predictability, de-risk appraisal well locations and to better understand producer-injector connectivity.
The study field is located in the Western Desert, Onshore Egypt and comprises of Upper Cretaceous tidal channel systems across four key reservoir levels where sand thicknesses range from 2 to 15 m. The field was discovered in 1993 but development drilling only commenced in 2008. The last integrated field study was performed in 2012. The analysis of wells drilled post-2012 indicated that there is a considerable depth difference along the flanks of the structure between seismic predicted depths and actual well tops (>50 m). The fault interpretation also required a re-look so as to reduce the lateral uncertainty of the main boundary fault and explain the lack of injection response in some areas of the field. This necessitated an update of seismic interpretation, static and dynamic models. A new interpretation could help identify attic volume upsides and help mature new appraisal and producer-injector locations. Further work was also proposed to test the feasibility of using seismic inversion for facies discrimination.
The available Pre-Stack Depth Migration (PreSDM) data was re-interpreted as part of the project. The fault interpretations were quality checked using Semblance/Dip maps, sand box models and wherever possible, were tied to the fault cuts seen in previously drilled wells. The time horizon correlation and seismic polarity were verified and were also cross-checked with the P-Impedance volume before being used in the static modelling workflow. The PreSDM Interval velocity model was used for depth conversion, where an anisotropy correction was applied to tie the wells. Vok and Polynomial methods were also applied, which in turn were used to derive depth uncertainty estimates. The update in the main bounding fault interpretation generated new appraisal locations in the deeper levels. The new interpretation was tested against the results from the latest drilling campaign in the field, and nine out of ten wells were within the one standard deviation uncertainty range.
Simultaneous inversion of the seismic data was also carried out as part of the project using the acoustic, shear and density data from 6 wells over the field. The inverted P-Impedance and S-Impedance were converted to Net to Gross (NtG), and were checked against the remaining 24 wells, which helped in validating the property cubes.
Forward wedge modelling suggested that individual sands of less than 15 m thickness would not be resolved from seismic due to seismic bandwidth limitations. Still, a review of inversion data together with geological insights and dynamic data helped to identify the high NtG areas across the reservoirs.
The integrated interpretation of inverted volumes with well and production data resulted in new insights into the field and helped to mature new appraisal and development well locations.
The study area is located in the Carnarvon Basin, offshore Western Australia. The paper discusses an approach for predicting the lateral variation of Net-to-Gross using 3D probabilistic seismic inversion. The goal is to define and understand the distribution of sands and shales on the basis of seismic reflection data. This modelling and inversion is supported by the good quality of seismic data. The seismic amplitude patterns indicate hydrocarbon presence and variation in properties like Net-to-Gross.
Qualitative 3D CtL ("Check the Loop") is used to highlight the mismatches between the forward modelled synthetic data generated from the model and the actual seismic. The possible causes of these mismatches are investigated and suggestions are made to improve the model in the key areas of mismatch. The Quantitative 3D CtL is used to further constrain the reservoir model by the seismic data. A 3D probabilistic seismic inversion for the main uncertain reservoir parameter using Shell's proprietary stochastic model-based elastic inversion algorithm is deployed.
After the 3D CtL with the static reservoir model, reserves in both North and South blocks of the field were reduced, whereas the probabilistic inversion highlighted the presence of a high Net-to-Gross (NtG) channelized feature which facilitated an additional volume booking for the central block. These insights immediately impacted development decisions.
This study underpins the benefits of seismically constrained reservoir modelling. The use of probabilistic inversion to map out geological features (like higher NtG channels) is a new insight in the applicability of this methodology.
Goudswaard, Jeroen (Centre for Technical Geoscience, Laboratory of Seismics and Acoustics, Delft University of Technology) | Dillen, Menno (Centre for Technical Geoscience, Laboratory of Seismics and Acoustics, Delft University of Technology) | Wapenaar, Kees (Centre for Technical Geoscience, Laboratory of Seismics and Acoustics, Delft University of Technology)
In this paper a procedure is proposed for multi-angle imaging and elastic multiscale involved are: (1) a decomposition into up and downgoing P- and S -waves; (2) compensation for the P- and S - downward propagation effects (inverse extrapolation); (3) imaging of the angle-dependent P -P and P -S reflectivity; (4) multiscale analysis of the imaged angle-dependent P-P and P -S data. It is shown that in the synthetic case, proper use of walkaway VSP data can improve the quality of the angle dependent depth images, because internal multiples can be effectively suppressed when images of all receivers are combined. With real VSP data we will have an additional improvement, caused by the fact that the receivers are closer to the interfaces, which will decrease propagation related disturbances of the reflected wave-field.
No preview is available for this paper.
Goudswaard, Jeroen (Centre for Technical Geoscience, Laboratory of Seismics and Acoustics, Delft University of Technology) | Wapenaar, Kees (Centre for Technical Geoscience, Laboratory of Seismics and Acoustics, Delft University of Technology)
No preview is available for this paper.