Inversion for reservoir pressure change using overburden strain measurements determined from 4D seismic
Abstract
When significant pore pressure changes occur because of production from a
hydrocarbon reservoir the rocks both inside and outside of the reservoir deform.
This deformation results in traveltime changes between reflection events on timelapse
seismic data, because the distance between reflection events is altered and the
seismic velocity changes with the strain. These traveltime differences are referred to
as time-lapse time shifts.
In this thesis, time-lapse time shifts observed in the overburden are used as an input
to a linear inversion for reservoir pressure. Measurements from the overburden are
used because, in general, time shift estimates are more stable, the strain deformations
can be considered linear, and fluid effects are negligible, compared to the reservoirlevel
signal.
A critical examination of methods currently available to measure time-lapse time
shifts is offered. It is found that available methods are most accurate when the time
shifts are slowly varying with pressure and changes in the seismic reflectivity are
negligible. While both of these conditions are generally met in the overburden they
are rarely met at reservoir level.
Next, a geomechanical model that linearly relates the overburden time-lapse time
shifts to reservoir pressure is considered. This model takes a semi-analytical
approach by numerical integration of a nucleus of strain in a homogeneous poroelastic
halfspace. Although this model has the potentially limiting assumption of a
homogenous medium, it allows for reservoirs of arbitrary geometries, and, in
contrast to the complex numerical approaches, it is simple to parameterise and
compututationally efficient.
This model is used to create a linear inversion scheme which is first tested on synthetic
data output from complex finite-element model. Despite the simplifications of the
i
inversion operator the pressure change is recovered to within ±10% normalised error
of the true pressure distribution.
Next, the inversion scheme is applied to two real data cases in different geological
settings. First to a sector of the Valhall Field, a compacting chalk reservoir in the
Norwegian Sea, and then the Genesis Field, a stacked turbidite in the Gulf of Mexico.
In both cases the results give good qualitative matches to existing reservoir simulator
estimates of compaction or pressure depletion. It is possible that updating of the
simulation model may be assisted by these results. Further avenues of investigation
are proposed to test the robustness of the simplified geomechanical approach in the
presence of more complex geomechanical features such as faults and strong material
contrasts.