An effective method of stochastic simulation of complex large-scale transport processes in naturally fractured reservoirs

Access full-text files

Date

2002-05

Authors

Hu, Yujie

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Increasing attention is being paid to uncertainties in reservoir production predictions because of the impact on business risk. The main concern in transferring uncertainty from reservoir simulation inputs to outputs is feasibility in terms of computing time and cost, due to the very large and complex subsurface geological systems. This study primarily focuses on developing an effective, accurate and robust method for estimating uncertain reservoir simulation predictions. Large amounts of uncertain simulation inputs are taken into account in the uncertainty study. Geostatistical and Monte Carlo techniques are used to sample uncertain data to establish an arbitrarily large and statistically significant collection of data sets. A naturally fractured reservoir model with a large amount of uncertain spatial reservoir properties is used for the study. The model includes such vii spatially varying reservoir properties as anisotropic matrix and fracture permeability, matrix and fracture porosity, relative permeability curves, capillary pressure curves and matrix block dimensions. Recovery is by water injection. Reservoir simulation uses dual porosity model. Two fundamental processes are used for the study. First, the results of uncertain matrix properties are examined. The equivalent matrix system concept, based on single-matrix-block imbibition studies, is used to effectively estimate uncertainty transferred from uncertain matrix property inputs. Probability rank-preserving properties are further examined between uncertain fracture properties and reservoir simulation outputs for the fracture system. Uncertain matrix properties are then incorporated with uncertain fracture system properties to complete the study. A Monte Carlo technique is used to provide true uncertainty results for comparison with estimation results. A reasonable method for estimating uncertain simulation outputs is given for waterflood recovery from a complex naturally fractured reservoir.

Description

text

LCSH Subject Headings

Citation