Prediction of Permeability from a Combination of Mercury Injection and Pore Image Analysis Data

A series-parallel (simplified network) model has been developed for which invasion percolation calculations modelling mercury injection can be carried out analytically. A considerable strength of the series-parallel model is its inclusion of a variable pore connectedness since this has as strong an influence on penneability as does the pore size distribution. Pore sizes are assigned to the model according to a distribution obtained from digital image analysis and mercury injection calculations are carried out for various values of connectivity. Agreement between experimental and modelled mercury injection curves is optimised by choice of a connectivity parameter with which the model permeability is then calculated. The result has been found to be a good prediction of directly measured permeability for a wide range of rocks.