Open Computing Platform Due For Shared Earth Modeling

THE PETROTECHNICAL OPEN SOFTWARE CORP. shared earth modeling (SEM) project has developed real-life scenarios for application of SEM concepts.

Schlumberger/Geoquest, a project sponsor, assembled this description of a progression of high-level use cases for the life cycle of a field development.

A new area is targeted for prospect analysis, and a 3D seismic survey is shot spanning several square kilometers. There is no prior drilling activity except for some wells in remote leases several kilometers away.

Part of the survey shows some prospect potential, and the 3D seismic is interpreted for several key horizons covering that area. Only faults that are clearly visible are interpreted, although other anomalies exist which may also be faults.

Interpreted horizon and fault structure boundaries are modeled in time within the selected prospect area extent, and structures within the SEM are defined. Estimates of rock properties (velocities, porosity, etc.) are made from data from the remote wells and correlation with the geology and seismic of the new region.

From a coarse velocity model based on these estimates, the structural model is converted into the depth domain. The depth-converted structural model allows a volumetric analysis of the prospect. 2D and 3D maps of the model are documented along with other supporting data, all of which are presented for final determination of the prospect.

It has been decided to go ahead with planning for up to three discovery wells for the Case A project. Several months have passed since the work on Case A was completed since time was needed to present to management and other investors.

The Case A project data set is recalled without any further changes to it and used as the basis for locating the trajectories of three wells. Some of the faults appear to seal parts of the formation and are critical to the process. The targets are often identified on the background of seismic information in time and subsequently depth-converted for the trajectory planning.

Well planning is completed, and the well paths and supporting data are added as part of the project.

The first well of Case B was a hit so a new effort is begun to register the discovery with the necessary regulatory agencies, modify planning of all subsequent wells, etc. The new well brings with it significant new data, including a checkshot survey.

From this new data, the Case A project is updated, the old structural models in depth domain are discarded, and new ones are generated which honor the velocity data from the checkshot. This velocity information is carefully extrapolated, and a new depth conversion is performed. Based on this depth model the volumetrics are rerun. In this way, all aspects of the project are upgraded to reflect the discovery well data, including replanning of the other two wells.

Updated maps and data are presented to outside regulatory agencies to set up field rules (pooling acreage, drilling densities, etc.). This phase of the project is archived as the discovery version and will be retained intact for legal and other reasons.

In this post-discovery phase, as new wells are drilled, the project database is continuously updated to support well planning, drilling, and production activities, and these updates are done on a clone of the discovery version.

Any intermediate versions are kept as snapshots only at key milestones in field development as required by management or for legal purposes. Where necessary, the layer velocity properties are updated, and the depth models are locally corrected to honor the effective depths of the drilled events. Reservoir characterization is kept as current as possible with all known data by continuously updating a field development version of the project.

Structural models are periodically rerun to update the SEM. The arrival of checkshot survey and other new sources of velocity data supports a higher-quality mapping from time to depth domain. This iterative procedure should be tracked in some meaningful way, possibly by versioning.

To optimize well production and field drainage, gas injection wells are introduced. Locating these wells requires a flow-simulation analysis of the reservoir, which it turn requires a 3D model of key reservoir properties.

The SEM must contain all the necessary structural boundary elements required to drive 3D property modeling. The SEM is updated to include the 3D property distribution, then used to drive flow simulation.

The usefulness of a time-lapse 3D seismic survey is investigated by sensitivity analyses.

The preproduction well logs and well logs obtained after some production time are used to estimate the effective parameters to be used in the fluid-substitution modeling. Synthetic seismic responses are modeled. The observed changes were regarded large enough to be visible through surface seismics.

This sensitivity analysis is also extended to a simulated time model in the SEM. The theoretical changes are mapped, and the areas where the expected effect is above a given threshold are considered as 3D time-lapse seismic targets.

A new 3D seismic survey is shot. The surface installations, which were not present in the preproduction survey, caused illumination and coverage problems in the time-lapse survey. The new survey is carefully calibrated by correlation of known events. A correction field is developed and applied to the entire seismic survey in the SEM. The calibrated time lapse and the original seismic survey are now correlated.

Differences are attributed to the model components where production effects originate (parts of layers in the SEM). The results are compared to the SEM flow model. The flow model is subsequently updated to include the findings of the time-lapse survey. Undrained compartments are identified. They are targeted by some high-angle boreholes that efficiently drain a series of small reservoir compartments.

In another area, a gas reservoir is obscuring the lower parts of the reservoir. The deployment of permanent cables with 4-component sensors is considered. The SEM serves the necessary data for a sensitivity analysis by elastic modeling. The cable length can be optimized to the reservoir.

The first 4-component service is used to set up the SEM with elastic parameters (compressional [C] and shear [S] velocities). The conventional seismic time is complemented by P-S converted and pure S-S reflection travel times, effectively providing the SEM two additional time property models.