Upscaling of porous media
Geological modeling methods are able to build models of enormous complexity
with help from modern computers. When engineers wish to simulate fluid flow
through rocks, perhaps in connection with designing an oil recovery plan or
minimising the risk from burying some toxic substance, they need to simplify
the geological models. The reason for this is that flow simulation needs to
solve large systems of linear equations which requires additional computer
storage. They also need to step the flow model through many time levels. Thus
it is necessary to reduce the memory and cpu requirements.
A simple example of such a simplification is when a set of electrical
resistors, arranged in series, is replaced by a single resistor with a value
given by the arithmetic average of the original set. For resistors read
permeability, and for arithmetic average read harmonic average.
When we move to 2 or 3 dimensions, and are concerned with more complicated
physics, the problem of replacing a detailed model by a simpler model so that
the large scale behaviour of the system is honoured becomes very difficult, and
very interesting.
Upscaling is closely related to homogenization theory, a branch of mathematics
that studies, amongst other things, equivalent media in a rigorous way where
the ratio of the small length scale of the details to the scale of interest is
asymptotically zero.
People working in this area within OCIAM
are
For detailed information see
|