ABSTRACT

This poster depicts the workflow approach that was developed to generate accurate reservoir volumes for a complex geological setting in the West Cameron block in Gulf of Mexico. The gas bearing sand has been divided into four blocks bounded by faults and has been confirmed by drilling in two blocks. The approach leverages an integrated software environment for consistent interpretation using seismic and log data. Seismic to well ties was an important element in the process. Next, the seismic data and interpretation (faults and horizons) are converted to depth and corrected to marker interpretation. A fault framework is then built by using the fault planes and generating the fault termination hierarchy. This framework is then used with the seismic horizon and well marker interpretation to generate the 3-D structural model.

Figure 1. This figure shows a seismic horizon just below the prospective sand. Seismic amplitude attribute map is draped on top of the structure. The bright (white) zones bounded by black outlines define the gas distribution in the sand. Also shown are the four compartments and the major fault controlling the distribution. The five wells considered in the study are shown with Sw curve displayed.

The approach analyzes the indications on the seismic attributes and validates them with log data. Several seismic attributes are analyzed and finally two attributes are chosen to estimate the log property from the seismic. Next, the geostatistical methods are leveraged to populate the framework using the estimated properties and the available logs data.

Figure 2. Original fault interpretation converted to depth is necessary for a consistent fault framework. The distribution of hydrocarbon in the field is controlled by the faults (cf. Fig 1), thus a good fault model is a key element in building geological model and computing correct volumetrics.

Subsequently, the sand object is defined using the voxel technology with depth converted seismic data. This sand object is then selectively populated using the best suited geostatistical technique. The approach was carefully designed to allow multiple scenario testing to optimize sand object definition. A high quality volumetric results were obtained by testing several hypotheses.

Figure 3. This picture shows the distribution of the prospective sand in Block 2 (cf. Fig. 1). The distribution is based on depth converted seismic data. The voxel picking parameters were defined at one well location and can be verified at the other well location. Several hypotheses for this sand object were tested by way of property distribution and volume calculations to decide the best parameter set for the sand. Finally the same parameters were used for picking in all the other blocks.

This optimized workflow allowed leveraging the available tools to get the optimal results.

This figure shows the Water Saturation distribution in the reservoir unit containing the prospective sand. Hot colors show higher gas saturation. Voxel picking is then used to define the sand object which is modeled and then selectively populated. This 3-D model can as-such be up-scaled for simulation.

End_of_Record - Last_Page 11-------