Print this pageMudstone (Shale!) 
Permeability
: A Key Unknown in Basin
Modelling
APLIN, ANDREW O., YUNLAI YANG, STEVE R. LARTER, DAVID N. DEWHURST, and JEAN-PAUL SARDA
The utility of fluid flow modules within basin models are currently severely
limited by a lack of basic data for shale permeability. Porosity:permeability
functions for "standard shales" within basin models give a permeability range of
four orders of magnitude at a porosity of 10 percent;
published-porosity:permeability data for shales define a permeability range of 3
orders of magnitude at a single porosity. One way forward is to recognise the
immense variability of shales and to realise that the rates at which shales lose
porosity and permeability during burial are strongly influenced by shale
lithology, defined for example by grain size (e.g. percent particles <2
microns). Our work has shown that both measured and modelled permeabilities of
shales with 30-40 percent <2 microns particles are up to 3 orders of magnitude
greater than those of shales with 60-75 percent <2 micron particles. Since shale
permeability is strongly influenced by porosity and lithology, we have been
looking at ways of rapidly determining shale lithology from wireline logs,
rather than just assuming shale properties! We have used data from shales
deposited on the Norwegian Margin to train Artificial Neural Networks (ANN) to
determine some physical properties of shales, including grain size. The ANN can
determine the percent <2 micron particles to within +/-5 percent of the true
value and should in the future be used in conjunction with better defined
porosity:permeability equations for individual shale lithologies to place much
better constraints on the shale-porosity:permeability
:effective-stress/depth
relationships used in basin models.