This paper assesses methods to predict the saturated hydraulic conductivity, k, of clean sand and gravel. Currently, in engineering, the most widely used predictive methods are those of Hazen and the Naval Facilities Engineering Command (NAVFAC). This paper shows how the Hazen equation, which is valid only for loose packing when the porosity, n, is close to its maximum value, can be extended to any value of n the soil can take when its maximum value of n is known. The resulting extended Hazen equation is compared with the single equation that summarizes the NAVFAC chart. The predictive capacity of the two equations is assessed using published laboratory data for homogenized sand and gravel specimens, with an effective diameter d10 between 0.13 and 1.98 mm and a void ratio e between 0.4 and 1.5. A new equation is proposed, based on a best fit equation in a graph of the logarithm of measured k versus the logarithm of d102e3/(1 + e). The distribution curves of the differences log(measured k) log(predicted k) have mean values of 0.07, 0.21, and 0.00 for the extended Hazen, NAVFAC, and new equations, respectively, with standard deviations of 0.23, 0.36, and 0.10, respectively. Using the values of d10 and e, the new equation predicts a k value usually between 0.5 and 2.0 times the measured k value for the considered data. It is shown that the predictive capacity of this new equation may be extended to natural nonplastic silty soils, but not to crushed soils or plastic silty soils. The paper discusses several factors affecting the inaccuracy of predictions and laboratory test results.Key words: permeability, sand, prediction, porosity, gradation curve.
Negative correlation between porosity and hydraulic conductivity in sand-and-gravel aquifers at Cape Cod, Massachusetts, USA
