ABSTRACT
One of purposes of reservoir characterization is to delineate reservoir heterogeneities as truly and effectively as possible. Reservoir heterogeneities are one of main factors to affect recovery ratio. Understanding of reservoir heterogeneities is useful to recognition of effects of barrier layers on fluid flow in reservoir, to clear seeing of spatial distributions of sandbodies and permeability, and to correct evaluation of spatial distribution of residual oil[1].
A number of studies shows that stochastic modelling and stochastic simulation of reservoir with geostastistics is an effective method to evaluate reservoir heterogeneities quantitatively[2][3][4].
Reservoir heterogeneities exist on various observation scales of reservoir, and full into large-scale heterogeneities and micro-scale heterogeneities. Different heterogeneities with different scales restrict and affect each other. More attention has been paid to stochastic modeling of sedimentary facies. The reason is that spatial distribution of sandbodies or geometry of sandbodies is important to control reservoir behaviour, to make development plan, and to estimate reliable recovery ratio of oil and gas [5] [6] [7].
Multi-stage stochastic modelling of reservoir, as an advanced method for reservoir characterization, can show various heterogeneities with different scales, and embody constraint relationships among these heterogeneities. It can make simulation of spatial distribution of sedimentary facies match with simulation of spatial distribution of petrophysical parameters. The process and results with multi-stage stochastic modeling will have more reasonable geological meaning than other stochastic modeling
The systems for reservoir characterization, HERESIM developed by IFP and Geostatistical Center of Paris Mining Institute, and MOHERES developed by Normay State Oil and Norway Computing Center, both have used two-stage stochastic modeling.
This paper proposes a methed for three-stage stochastic modeling of reservoir with fluvial facies. This method consists of three models: stratification plane configuration model, sandbody framework model, and petrophysical parameter distribution modeL A case study, presented in this paper, indicates that this method is especially suitable for fine quantitative description of complex fault block reservoir with fluvial facies.
