Rock fabric, pore geometry and mineralogy effects on water transport in fractured dolostones

Abstract

The combined influence of rock fabric, pore geometry and mineralogy (petrological parameters) on transport properties in complex and heterogeneous naturally fractured rocks was studied experimentally. As fissure geometry quantification has rarely been addressed in most theoretical studies of transport properties, emphasis was placed on the effects of fissure geometry on both non-saturated media (capillary rise) and saturated media (permeability). We proved that prediction of transport properties in complex rocks is not guaranteed by the estimation of the classic micro-structural variables considered in the literature and that accurate prediction can only be attained when petrological parameters are first quantified in depth, and then combined. Principal component analysis and the regression models proposed here clearly demonstrated on the one hand that, a meaningful porous network in non-saturated media must be modelled with a combination of different geometrical capillary tubes representing the matrix (cylindrical) and the fissure (rectangular prism) and on the other hand, that in saturated media a well-in-deep fissure size quantification enabled a more accurate prediction of permeability to be made. The experimental data confirm that transport properties and its anisotropy are closely dependent on fissure typology, textural characteristics, mineralogy and spatial distribution of the whole rock fabric elements. Enlarged-fissures weakly exert capillary suction due to the retarding effect of gravitational forces, but they are vital in controlling permeability. Cracklebreccias with small clasts, high dolomite cement content and high inter-clast fissure density exhibit strong capillary suction. However, high calcite cement produces abnormally low rates of capillary rise, due to possible pore surface contamination, together with a high contact angle effect. Good agreement between permeability and geometric factors provided a suitable basis for identifying preferred permeable directions. Additionally, we found a critical fissure density which defined the isotropic matrix permeability. We also present a new practical and simple linear model relating permeability to capillarity with meaningfully and easily estimated petrological parameters. Results obtained in the present study demonstrated the correct identification and use of more directly related petrological variables for modelling transport properties. Moreover, the analysis of these results using multivariate analysis is considerably more demanding compared to the conventional approaches.