Modeling of a Tide-influenced Point-bar Heterogeneity Distribution and Impacts on Steam-assisted Gravity Drainage ProductionExample from Steepbank River, McMurray Formation, Canada
Résumé
Outcrops and subsurface investigations emphasize that the main bitumen reservoirs of the
McMurray Formation are large point-bar deposits. Sedimentological studies performed on
these reservoir objects have shown that tidal currents occurred in the meandering paleoriver
system. These tidal inputs increased reservoir heterogeneities primarily because of successive
mud decantation periods and the many reactivation or erosion surfaces.
Five main reservoir heterogeneities have been described on Steepbank River outcrops:
mud accumulation during channel abandonment, mud drapes along accretion surfaces that
are downward interfingered into cleaner sands, flood-plain deposits on top of the point bar,
reactivation surfaces typically associated with mud-clast deposits, and mud-clast breccias accumulated
at the base of the channel. At the same time, five main facies have been emphasized on
these studied tide-influenced point bars: mud-clast breccias, cross-stratified sands, slightly heterolithic
rippled sands, highly heterolithic burrowed sands, and thick mudstones. For each of these facies, petrophysical properties have been established, enabling their application as rock
types for the Steepbank River outcrop modeling. This deterministic method of modeling,
improved by light detection and ranging (LiDAR) data, used truncated Gaussian simulations
constrained by the proportion cube, sedimentological logs corresponding to hard data, and
adequate variograms. The resulting facies and heterogeneity distributions conform closely to
the outcrop reality, lending support for the modeling method.
This deterministic sedimentological model is at the root of steam-assisted gravity drainage
(SAGD) simulation tests performed for a water-steam injector horizontal well underlain by a
fluidized-bitumen horizontal producer well. Results of the simulation boxes filled by this sedimentological
modeling more closely match outcrop analogs than results obtained by the more
commonly used purely stochasticmodels. This is caused by the occurrence of large-scale inclined
heterolithic stratification (IHS), which is generally overlooked in stochastic models, or alternatively
is replaced by horizontal heterogeneities, which would be highly detrimental for
bitumen production from SAGD horizontal wells. Thus, in purely stochastic modeling, bitumen
recovery prospects tend to be underestimated, or permeability heterogeneity is overestimated.
Finally, a comparison was made between SAGD simulation tests performed in the direction
of point-bar lateral accretion versus tests performed parallel to the paleoriver flow direction.
The comparisons show stronger bitumen recovery for the SAGD simulation tests performed
parallel to the paleoriver thalweg. In this case, parallel to paleoriver flow direction, a better
conformance of the steam chamber along the trajectory of the horizontal injector well is observed;
and the steam chamber rises more slowly toward the ground surface. This is because in
the SAGD simulation tests performed in the direction of point-bar lateral accretion, major reservoir
heterogeneities exist as thin low-permeability mud beds within the large, low-angle, IHS.
Such mud bed reservoir heterogeneities are less prevalent, and less restrictive, in the direction
parallel to the paleoriver flow.