Post-Stack Inversion Predicts Tight Sand Reservoir Porosity

_

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 226856, “Tight Sand Reservoir Porosity Prediction Using Post-Stack Inversion,” by Zainaw Alelew, Somanath Misra, and Sultan Sayghe, Saudi Aramco, et al. The paper has not been peer-reviewed.

_

This paper focuses on the geophysical challenges encountered during reservoir characterization and quantitative interpretation phases, particularly with regard to porosity estimation. Porosity prediction becomes challenging in low-porosity or tight sand reservoirs as the propagating seismic waves become less sensitive to changes in porosity. Consequently, a novel approach is followed to predict reservoir porosity.

Objectives

This study presents the following three distinct workflows implemented as part of this project:

- Use of regional well data to generate a prognosis for the target reservoir, which subsequently is employed to project hypothetical well tops in wells that do not penetrate the target reservoir.

- Application of local corrections to seismic interpretation that removes velocity-pullup structures caused by the overlying channel system, thereby enhancing the accuracy of the interpretation.

- Implementation of post-stack inversion for porosity prediction, accompanied by a range of quality-control (QC) measures to ensure the reliability and validity of the results.

To maintain confidentiality, the target reservoir will be referred to as Formation A, while the overlying formation will be designated as Formation B.

Methodology

Data Preparation.

The seismic data used in this study consists of a reprocessed prestack time-migrated 3D volume. Before conducting seismic interpretation, a comprehensive data collection and QC process was undertaken, encompassing all relevant data, including well coordinates, well logs, check shots, and well tops. Subsequently, the boundary of the area of interest (AOI) was delineated by considering both the field’s boundaries and the extent of well-data coverage. Furthermore, key wells were identified for use in the post-stack inversion process, and a blind well was selected to be excluded from the full inversion and porosity-prediction workflow, thereby facilitating a blind test of the predicted results.

Seismic Interpretation. The initial step preceding horizon mapping involved tying the wells to the seismic data, thereby establishing a direct correlation between the seismic data in the time domain and the actual well data in the depth domain. This was achieved by utilizing P-velocity logs, density logs, and available time-depth relationships (check shots at either well location or from the nearest available well) to calculate the reflectivity and generate synthetic seismograms, which enable the seismic data to be tied to the existing well logs.

Three horizons, including the target reservoir, were identified and subsequently mapped manually (without snapping or extrapolation) across the area at 25×25 intervals. The interpretation process began with the selection of a crossline that connected all the wells within the AOI, and then the interpretation was extended to encompass the full extent of the area. The resulting interpretations were used to generate surfaces with a 50×50 grid increment, which would be used later for inversion purposes.