Design, specification and construction of filters - Geotechnical Engineering of Dams - page 356

Figure 9.13. Lafluer et al. (1993) method for design of conventional and geotextile filters.

criteria for aggregate filters is D 15 /5 d SF , so they have opted for coarser

filters than in the ICOLD (1994) and Lafleur et al. (1989).
Figure 9.13 does not include design criteria for cohesive soils. Lafleur et al. (1989) rec-
ommend D 15F

d SF , not D 15 /4

0.4 mm for non dispersive cohesive soils and D 15F 0.2 mm for disper-

sive soils.
Sherard and Dunnigan (1985, 1989) recommend, for their Group 3 soils, that D 15F

4D 85B , whereas Lafleur et al. 1993 recommend, for uniform granular soils, D 15F 5D 85B .

Given that Foster (1999), Foster and Fell (1999a) showed that erosion would continue for
D 15F

9D 85B (see Section 9.3 ), both criteria
seem to have some margin of safety against large erosion losses.

9D 95B and would be excessive for D 15F

9.2.1.6 Brauns and Witt
Brauns and Witt (1987), Schuler and Brauns (1993, 1997) and Witt (1993) propose that
filter criteria should be based on a probabilistic analysis to allow for the variability in the
particle size distributions of the base soil and the filter. Honjo and Vaneziano (1989) had
a similar concept. The authors can see some merit in these approaches, but they are pred-
icated implicitly that filters either “fail” *or “succeed”. As discussed in Section 9.3, the
reality is that such criteria apply to no-erosion acceptance criteria and that much coarser
filters will eventually seal.

9.2.1.7 Foster and Fell
Foster (1999) and Foster and Fell (1999a) carried out extensive no-erosion tests using a
test set-up similar to that of Sherard et al. (1984a,b), and reviewed the results of the
USSCS tests as reported in internal reports by Sherard (Sherard, 1985a and b).