The answer is going to depend very much on the sand itself and the depth of the fill - and on
how much you want to know the *rate* of settlement as well as the amount. It's comparatively
easy to use standard lab tests to identify the compressibility parameters so that you get
pretty close to the final thickness, but much more difficult to predict the time scale unless
the permeability is high and the depth of fill is low enough that the settlement occurs
quickly. The approach suggested by S W Lee, Geotechnical Research Office, Cambridge
University from the book, Soil Mechanics - Concepts and Applications by Prof. William
Powrie of Southampton University, pp. 149 - 151, assumes that the sand layer doesn't settle
very much, and it also assumes a constant and uniform soil stiffness. These assumptions may
indeed be perfectly acceptable for your particular problem - but you need to be aware of them.
One of my research students Gert Bartholomeeusen and I recently organised a prediction
symposium at Oxford for exactly the situation you describe - ie, a soil slurry settling under
its own weight. We invited specialists in large strain consolidation to predict the results
of a basic settling experiment, given the results of three other tests on the same material,
but with different initial densities and heights. The soil was not just sand - so your
situation may be less difficult than our sandy silt - but the general outcome was a range of
settlement prediction in which the largest was about 50% greater than the smallest - but there
were orders of magnitude difference in estimates of the time scale! Most of the participants
used the Gibson large strain equation as their starting point.
There are at least a couple of ways forward depending on your conditions. You can improve the
approach quoted by Powrie by using a large strain iteration (thus removing the constraint that
the settlement has to be small) as done, for example, by one of my research students in his
thesis (Lee, 1999). This doesn't change the need to assume unique and constant parameters to
describe the soil behaviour, but is simple and should give a better solution than the small
strain one.
Another possibility is to use one of the standard large strain solutions of Gibson's 1967
equations in which it is possible to take account of changing compressibility and permeability
with stress level and void ratio. The conclusion from our prediction symposium was that the
crucial decisions were the choice of parameters from the available data. (We noted, in
addition, that if the soil showed significant creep behaviour, even this theory was
inadequate! However, you've raised the question about a sand fill, so this is unlikely to be
relevant.)
Hope this helps
Regards
Gilliane Sills
On Tue, 19 Dec 2000, Alex Carson wrote:
> The following question was asked in the office, which I could not answer.
> Is there a reliable method of calculating the settlement and rate of settlement of hydraulic sand fill ?
> All answers will be gratefully recieved
>
> Alex Carson
> Principal Geotechnical Engineer
> Telford
>
>
----------------------
Dr Gilliane Sills
Department of Engineering Science
University of Oxford
Parks Road
Oxford OX1 3PJ
Tel (44) (0)1865 273164
Fax (44) (0)1865 273907
[log in to unmask]
|