Like most geoscience disciplines, time-lapse seismic (also known as 4-D seismic) is not an exact science. Feasibility studies from multiple disciplines must be taken into consideration and forward-modeled prior to the actual survey beginning.

This was the message from a paper presented at the 2015 Society of Exploration Geophysicists annual meeting by Kurt Eggenberger, David Hill and Dominic Lowden, formerly with Schlumberger, and current Schlumberger employees Sonika Sonika and Mehdi Paydayesh. The paper, “High-fidelity 4-D forward modeling as part of a redefined closed-loop seismic reservoir monitoring framework: A case study,” provides a way to model anticipated seismic responses in a more quantitative way through a closed-loop procedure.

“Literature produces many successful examples of time-lapse seismic case studies with carefully analyzed 4-D signatures,” the authors noted. “However, quantitative and even qualitative comparisons of differences between predicted and actually measured time-lapse seismic data are much less performed and discussed in open literature despite the obvious benefit for model reconciliation.

“An explanation can be found in the assumptions and methodology used for the feasibility study, which can be too simplistic to warrant … a meaningful comparison.”

The authors go on to note that the actual 4-D signatures often are larger, smaller or different in shape than what the feasibility study predicted. These differences can be expressed as a difference between the actual geology and the modeled geology, between the actual physical properties and the model properties or by not taking into account the full complexity and interactions.

Closed-loop model
While proposed as early as 2000, the closed-loop approach was considered too costly to be practical in 4-D studies. But as compute power grew, two approaches became more commonplace. One is a well-log-based fluid-substitution and petrophysical modeling approach, which neglects realistic noise, overburden and acquisition effects. The second is a simulation-to-seismic modeling approach that incorporates overburden and acquisition effects by ray-tracing. This method also neglects realistic noise, however.

A newer method uses 1-D convolutional modeling to create a cube for each angle of incidence and each vintage of data.

All of these approaches confine the modeling to the reservoir interval, the authors noted. Yet they’re widely used because they’re fast.

To truly model the 4-D response as it relates to production, the authors propose a methodology that incorporates the overburden, sideburden and underburden into the modeling. “To accurately predict a 4-D response related to production, modeling cannot be limited to the reservoir only and has to be extended to the full field and its geomechanical effects, describing the acoustic and elastic response to both reservoir and fieldwide changes,” they noted. This type of model must take into account the reservoir dynamics component encompassing fluid properties, fluid flow characteristics, field performance history, and pressure distributions and profiles over time. It also must incorporate pressure changes due to production. This, in turn, can only be accomplished by integrating geologic, reservoir stimulation and reservoir geomechanical models into a full Earth model.

Modern computing systems are now enabling fieldwide 3-D finite difference acoustic and elastic models to be viewed in conjunction with the elastic properties from the Earth model (Figure 1). At each step the modeled and measured data are analyzed and reconciled.

FIGURE 1. The redefined closed-loop seismic reservoir monitoring workflow caters to a 3-D dynamic integrated Earth model. White boxes indicate additional elements to earlier proposed frameworks. This workflow also allows model reconciliation based on high-fidelity seismic response. (Source: Schlumberger)

Chimera modeling
Schlumberger’s Chimera model has been built to test the closed-loop monitoring framework. Figure 2 depicts an offshore turbidite-type reservoir with 25% maximum porosity and 200 nD maximum permeability in 200 m (656 ft) of water. A four-way closure contains the hydrocarbon accumulation,
and a number of vertical faults are present.

To simulate the model, the authors generated time stamps in three-year intervals with the monitor baseline set in 2014. The reservoir was simulated in primary production for three years followed by waterflooding. Pressure, water, gas and oil saturation simulations were transformed through a petro-elastic rock physics model to get elastic properties.

Ancillary modeling parameters were then determined by 1-D conventional, ray trace and wedge modeling; then the finite-difference modeling was undertaken with the derived properties. This approach also requires a master geometry definition along with a representative noise model. Results indicated that this approach gave a better sense than conventional approaches of the amount of difference that could be expected between the modeled signature and the measured signature.

By incorporating geology, reservoir simulation and geomechanical models into an integrated full-field coupled dynamic integrated Earth model, the authors were able to derive high-quality elastic parameters by the petro-elastic rock physics model for input into the forward modeling. Also, by using finite-difference modeling with realistic calibrated noise, a high-fidelity prediction of the 4-D signal decreases modeling errors.

“Having modeled with such a high-fidelity 4-D signal, it can be determined if the baseline acquisition geometry will measure the 4-D signal at the required time interval and, if not, what acquisition geometry will,” the authors noted. “Furthermore, seismic reservoir monitoring can make use of the high-fidelity forward-modeled data that goes beyond the traditional feasibility and survey design study with the objective to reconcile modeled with actual measured data, closing the loop to a dynamic integrated Earth model. Higher fidelity predictions of future reservoir behaviors are the result.”

FIGURE 2. The Chimera reservoir model represents a faulted dome structure in a clastic environment governed by shale and sand. Figure 2a (left) is a full-field permeability model on the Earth grid, and Figure 2b represents the reservoir porosity model. (Source: Schlumberger)

Acknowledgment
This article is based on a presentation at the 2015 annual meeting of the Society of Exploration Geophysicists and has been written with permission by the authors. Kurt Eggenberger, David Hill, Dominic Lowden, Sonika Sonika and Mehdi Paydayesh (2015) High-fidelity 4-D forward modeling as part of a redefined closed-loop seismic reservoir monitoring framework: A case study. SEG Technical Program Expanded Abstracts 2015: pp. 5424-5429. doi: 10.1190/segam2015-5869039.1