Overestimation of soil´s hydraulic conductivity using the Kozeni-Carman permeability model

[abdiel_leon]

Dear Engineering professionals,

I have a question concerning the possible overestimation of hydraulic conductivity in a fine quartz sand (0.06 mm - 0.71 m, D50 = 0.23 mm, Cu = 1.8. I am performing a numerical simulation of mechanized excavations in a saturated soil. To this end, the KZ model gives an estimate of hydraulic conductivity of 2e-4 m/s. However, numerical results are close to experiments when K ~ 3e-5 m/s. I am doing my research in the field of numerical geomechanics and the experiments were not the strong point of this work, so no permeability measurments have been carried so far. Just as a general question, is it possible the overestimation of the real, effective soil permeability using this permeability model? and if so, then to what extend?

Thank you very much on any hint provided.

Best,
Abdiel

 

[LRJ]

Hydraulic conductivity (often erroneously referred to as 'permeability' by geotechnical engineers) is notoriously difficult to measure, despite its fundamental importance for many geotechnical analyses. Generally the Kozeny-Carman equation works well, but I seem to recall it falls apart a little when there is high fines content. The plasticity of the fines might even play a role. Does that apply for your case? It doesn't surprise me that hydraulic conductivity results could be out by an order of magnitude - indeed that's common even from lab and in situ testing.

The other aspect here is simply that your constitutive model in your numerical analysis doesn't actually describe the world of geotechnics as well as you think. What makes you think the hydraulic conductivity is the source of the error? Have you plotted your single element stress-strain curves to see how well they match laboratory element tests? Have you undertaken sensitivity analyses to assess the bounds of other input parameters to see whether results can match better if those other parameters are tweaked?