Right click on poster to save.
PSPrediction of Sub-seismic Sealing Faults Using Simple Numerical Simulation Models*
R.C. Bain1, K.H. MacIvor1, B.E. Holt1, and D.S. Beaty1
Search and Discovery Article #40242
Posted May 28, 2007
*Adapted from poster presentation at AAPG Annual Convention, Long Beach, California, April 1-4, 2007
1Chevron North America Upstream, Houston, Texas ([email protected])
In order to justify development drilling in a partly-depleted, highly faulted gas reservoir in which untapped higher-pressure compartments may exist, convincing evidence for fault separation from existing producing wells must be provided, either by obvious fault breaks on 3-D seismic or by missing section due to a fault encountered in a well. Lacking such evidence, it is difficult to state with certainty that prospect reserves will be incremental, as opposed to acceleration, even when volumetric analysis suggests that existing wells will not capture all of the producible reserves in a reservoir.
The Mid-Continent Business Unit of Chevron North America Exploration and Production has had success in the Lobo Trend of Webb and Zapata Counties, South Texas, using simple, "fit-for-purpose" 3D-earth models and numerical simulation models that provide a level of confidence sufficient to predict the location and expected reservoir conditions of remaining incremental reserves in a partly-developed reservoir. These models have proven to be very useful in their ability to provide quick results with limited geologic and reservoir data. The key factor in their success is the proper integration of flowing pressure data with observed production decline curves. Static reservoir pressure measurements are typically unavailable and also give misleading results when used for P/Z volumetric analysis in compartmentalized reservoirs.
In the first example, a simple simulation model predicted the presence of sub-seismic faulting that provided a seal for the objective reservoir. The proposed location was in a syncline between two wells that had already produced large volumes of gas and were producing at very low bottom-hole pressures. An iterative approach involving the seismic interpreter and the reservoir engineer resulted in a geologic model that was supported by the seismic data and agreed with the history matching efforts. The well, which would not have been approved without the model to support it, encountered near-virgin reservoir conditions.
The second example provides a lesson learned, demonstrating a reservoir in which the reservoir simulation and history match correctly predicted the presence of a sealing fault, but incorrectly predicted which of several faults was the sealing one. The sealing fault was penetrated by the wellbore and the seal was ruptured when the well was fracture stimulated.
Quickly demonstrating the accuracy and applicability of simple numerical models in an environment where rig moves are rapid and reservoir data is sparse has generated a new interest in a tool that was heretofore thought too complex and too time consuming to apply. Asset Team Earth Scientists are now working more closely with the reservoir simulation engineers and are using the results from these simple models to help in their interpretation of subsurface geology, especially in highly faulted environments. In some cases, successful wells are being drilled where they otherwise would not have been.
Poster 1: The Problem, Geologic Setting, and Reservoir Simulation Basics
Volumetric calculations indicate that two wells producing from the same gas reservoir have not drained all of the producible reserves in a 200-acre fault block. The challenge is to identify economic drill locations despite the fact that the existing wells appear to be depleted.
Posters 2 and 3: Case Study #1
A 3-D seismic interpretation of the fault block described above was converted to a GOCAD model and assigned reservoir properties. A 3-D cellular model was then constructed. History matching of the actual well performance to the numerical simulation indicated that internal boundaries were required to adequately explain the pressure and rate performance of the wells. This caused the seismic interpreter to revise the 3-D model.
By interpreting tiny offsets of flexures in seismic events as possible faults, the seismic interpreter was able to segment the objective reservoir into four blocks (see Line A-A’). The 3-D cellular model was revised to incorporate the new barriers, and the simulation was run again. This time the pressure and production history of the wells matched the simulation model closely. This provided the confidence necessary for proposing additional wells to fully develop the block. Prior to the revised interpretation, the presence of the existing, nearly depleted wells in the block would have discouraged any additional drilling. Results of two new wells demonstrate that the revised interpretation was correct. This 3-D interpretation would not have been made without the results of the simulation, which suggested that additional faulting was present.
Poster 4: Case Study #2
In a different part of the model created in Case Study #1, a well was proposed to offset a competitive drainage situation in a 60-acre fault block. The proposed well was expected to encounter similar pressure to a recently drilled updip well. An unexpected 80-foot fault was encountered in the objective reservoir when the well was drilled, but the pressure matched what had been predicted by the simulation model. However, after fracture stimulating the well, the observed pressure had decreased by more than 2000 psi. The fault encountered in the well was apparently a seal between two compartments of vastly different pressure. The stimulation ruptured the seal and the frac job “went south.”
Flowing tubing pressure (converted to bottom hole flowing pressure) can be used for the pressure match when other reservoir pressure measurements are unavailable