LONGITUDINAL DISPERSION IN POROUS-MEDIA WITH RANDOM VARIABLE POROSITY

A solute dispersion model with random variable porosity is considered in the present study. A spectrum of longitudinal random porosities is generated by the random number generator to illustrate the applications and to assess its effect on the solute dispersion. The complicated governing dispersion model can be conveniently tackled by an implicit finite difference method. It is shown that if the random porosity function does not deviate more than 92, 70 and 34% from its average, the dispersion model using average porosity can, respectively, predict an average solute concentration within 15, 10 and 5% of that of the random variable porosity model.