Print this pageAbstract: Integrating Deterministic and Stochastic Reservoir Description for Simulation
A combination of deterministic and stochastic techniques are used to develop a methodology for generating descriptions for a variety of simulation models. The models range from site specific to generic and/or mechanistic.
Deterministic components are those which are mappable and predictably continuous over many well spacings, such as structure, zone thickness, and large shales. However, we are conservative in what is included here in order to avoid forcing unwarranted (and potentially time-consuming) "deterministic" interpolations, and so that we can take advantage of the capability of stochastic methods to rapidly investigate uncertainties/sensitivities.
Stochastic components are porosity, permeability, and small shales, which are not correlatable at normal well spacing. A range of geostatistical and/or object-based techniques are used to simultaneously create fine-scale models of all flow-related properties. These models are then re-scaled to the simulation grid size, avoiding the application of inappropriate core-derived correlations.
It is important to clearly define the stochastic and deterministic strands at the outset, so that both can proceed in parallel, ensuring that all aspects are covered but not replicated. The models are completed by merging the deterministic maps with the re-scaled stochastic properties.
We argue that it is important for geoscientists to be involved at an early stage in simulation studies since they can, and should, have considerable influence in establishing optimal model specifications. Also, we should be cautious with respect to the veracity of deterministic correlations if we are to capture the true range of uncertainty.
AAPG Search and Discovery Article #90986©1994 AAPG Annual Convention, Denver, Colorado, June 12-15, 1994