Grading curve relations for saturated hydraulic conductivity of granular materials

Estimation of hydraulic conductivity in soils is challenging. The primary aim of this study is to demonstrate that such predictions may be improved if grading curves are appropriately quantified and described, as well as by including density-related values in such relationships. Various saturated hydraulic conductivity models were tested with the assumption that predictions would improve if different grading curve statistics are used. A unimodal database was elaborated using old and new data. Three types of permeability models were examined. One using the traditional variables consisting of the product of harmonic mean d(h) or d(10) and void ratio, the hydraulic radius; as well as additional density information. The second using the grading entropy coordinate pair S-0, Delta S or the similar pair d(10), C-U, expressing the mean grain size on logarithmic scale along with the spread of the grain size distribution and containing similar information on pore size distribution (POSD) by duality. When these were combined in the third type, including also relative density for coarse materials, the fit was the best, verifying the hypothesis that the full pore size range may be the missing pore geometry information of the Taylor's equation (hence predictions are better if grading curve parameters consider the entire distribution of particle sizes). The parameters identified for the various data series were dependent on the data themselves as found from early times in literature. The similarity of grading entropy coordinate pairs and the pair d(10), C-U, as well as d(h) and d(10), was analysed by simulations and by using the same measured data.