Environmental Engineering Reference
In-Depth Information
As they point out, there are two main approaches:
(a) Uncoupled analysis, where the first step is to model the pore pressures using partially
saturated flow equations, followed by the use of limit equilibrium analysis. These
analyses allow, with varying degrees of sophistication, for the dependence of the
unsaturated permeability with the water content, stresses, degree of saturation and
diffusivity (equivalent to the coefficient of consolidation in saturated soils). The many
shortcomings of such analyses are detailed by Fry et al. (1996).
(b) Coupled analysis models (again with varying degrees of sophistication) the pore air
and water pressures, volume change and stresses using finite element analyses and
properties based on what are usually time consuming and sophisticated laboratory
tests. Fry et al. (1996) provide a review of the approaches available at that time.
Now (2003) there is considerable research in progress on this topic which should pro-
vide better methods. In view of this and the reality that the science of partially saturated
soils is well beyond most dam engineers (and the authors!), if it is wished to attempt a
fully coupled approach, one of the handful of experts in this field should be engaged to
assist.
11.4.2.5
Undrained strength analysis
In view of the difficulty in correctly estimating the pore pressures in partially saturated soils,
for most dams it is sufficient to use the undrained strengths (for cohesive soils) and a total
stress analysis.
The undrained strengths used are usually those estimated to apply at the placement
condition, so any strength gained by (partially saturated) consolidation as the dam is con-
structed are ignored.
Alternatively, the undrained strengths can be determined by a series of laboratory tests
which consolidate the samples to the effective stresses expected in the dam. These stresses
should be estimated by numerical analysis and the testing should follow the stress path of
the soil from compaction through to the conditions in the dam.
Just as there are uncertainties in estimating the pore pressures, there are uncertainties in
these undrained strengths.
11.4.2.6
Summing up
It should be recognised that it is not practicable to estimate construction pore pressures
accurately and that the best one can do is to get an indication of the likely pressures, and
how they will build up as the embankment is constructed. It is necessary to monitor pore
pressures with piezometers if it is considered they may be critical. Measures can be taken
to reduce pore pressures, e.g. by using a lower water content. This was done at Dartmouth
Dam (Maver et al., 1978). However by reducing the compaction water content the earth-
fill will be more brittle and more susceptible to cracking.
It is also important to monitor displacements of the dam during construction, so the
onset of increasing rates of movement, possibly indicating construction instability, can be
detected. Surface survey should always be possible. Bore hole inclinometers are valuable
but some, including the authors, are reluctant to put too much instrumentation into dam
cores because of the increased chances of piping problems along the instruments.
The discussion above, with the exception of Section 11.4.2.4, ignores the effects of dis-
sipation of pore pressure with time in the embankment. Eisenstein and Naylor in ICOLD
(1986a) discuss how this can be modelled by finite element methods. In most cases the
earthfill will have a low permeability and coefficient of consolidation and drainage paths
will be long which will result in long times for pore pressure dissipation. As the estimation
of pore pressures is approximate, it would not be warranted to take account of pore pres-
sure dissipation in most cases.
