(1) ARCO Technology and Operations Services, Plano, TX
(2) ARCO Technology and Operations Services, TX
Abstract: Capturing the impact of reservoir architecture on fluid flow: Deterministic versus stochastic models based upon outcrop analogs
Outcrop data from two depositional environments were used to evaluate the impact of reservoir architecture on waterflood simulation. The two environments are lower shoreface and fluvial meanderbelt. Two different models were constructed for each environment. A deterministic model was created by correlating logs in a “layer-cake” fashion. A stochastic model was created using discrete sand and shale objects whose geometry and abundance are related to the outcrops.
For the lower shoreface, waterflood simulation resulted in ultimate recoveries that differed by less than 1% OOIP at 1.0 HPVI. However, at any given time during production, the stochastic model implied that oil recovery was about 10% less than that of the deterministic model. This was due to a lower injection rate in the stochastic model caused by greater flow tortuosity associated with shale baffles contained within amalgamated sandstones.
For the fluvial meanderbelt, waterflood simulation of the deterministic model showed a 7% greater recovery for injector / producer pairs aligned parallel to depositional dip due to higher connectivity in this direction. The stochastic model showed 2% greater recovery for injector / producer pairs aligned parallel to depositional strike because of greater connectivity and higher flow tortuosity causing a larger pore volume to be swept. Although ultimate recoveries were similar in the dip-oriented deterministic and strike-oriented stochastic models, peak rates predicted by the deterministic model were 15 to 20% higher.
This study shows that inclusion of architectures derived from outcrops have significant impacts on rate and well placement that should be captured in simulations.
AAPG Search and Discovery Article #90914©2000 AAPG Annual Convention, New Orleans, Louisiana