new approach for approximation of dispersivity in porous media

this paper presents a simple yet powerful original approach for approximating dispersivity in porous media. the key to this approach is both use of least-squares criterion and image processing technique. two dimensional transparent models, including horizontal and vertical configurations, under different combinations of grain sizes and pore velocities, experimented. concentration data were obtained on a 0.5×0.5 mm grid, which represents approximately 1.8×105 measuring points. estimates of the longitudinal and transverse dispersivity are obtained for each experiment in three time step, and discussed. acquired results revealed that presented procedure successfully determine dispersivity coefficients and the maximum error of estimation is 4 to 7%. furthermore, calculated disperssivity varied from 0.001 to 0.003 m corresponding to grain sizes that agree with literature.