Assessment and
Reduction of Uncertainty in Fault
Seal Analysis for Hydrocarbon Fluid Flow
Prediction
Jones, Philip A.1, Stephen R.
Freeman2, Simon D. Harris2, Rob J. Knipe2,
Raoul R.D. Treverton2, P. William Bradbury2 (1) Rock
Deformation Research Limited, Leeds, United Kingdom (2) Rock Deformation
Research Limited,
A new technique for assessing and
capturing uncertainty in fault
seal analysis of hydrocarbon reservoirs is
presented. This allows for the rapid evaluation and modeling of uncertainty for
cross-
fault
fluid flow (e.g. via transmissibility multipliers). By utilising
the
fault
property visualisation techniques presented, only a small number of
scenarios need to be taken through to full reservoir flow simulation.
The natural variability of the input data
and limited availability of local datasets often leads to the introduction of
uncertain values or properties into fault
seal workflows. The effect of these
uncertainties on the results of a
fault
seal analysis is integral to the
workflow presented here.
Measured fault
rock property values from
core samples are used to calibrate
fault
rock clay mixing/smearing and
fault
permeability predictors under for application to structural and stratigraphic
models. Inaccuracies and uncertainties in both the input data and
fault
property prediction algorithms are incorporated. The ranking of probabilistic
scenarios allows an improved understanding of the likely impact of reservoir
cross-
fault
fluid flow, whilst at the same time allowing assessment of possible
risk factors.
The fluid flow characteristics are
assessed in terms of the fault
hydraulic resistance and the effective
cross-
fault
transmissibility (incorporating host cell properties). The impact
of these different cross-
fault
fluid flow property distributions are presented
by using streamlines. This visualisation technique allows for the rapid
comparison of meaningful properties that can be compared across multiple
uncertainty realisation models.
The rapid evaluation of multiple
uncertainty realisation models means only the most appropriate, fully risked
models need to be passed through to a full flow simulation.
AAPG Search and Discover Article #90063©2007 AAPG Annual Convention, Long Beach, California