Environmental Engineering Reference
In-Depth Information
Figure 11.26.
Eppalock Dam - Observed crest settlement (courtesy of Goulburn Murray Water).
density ratio (standard compaction). A weighting berm was built on the upstream side
from waste rock and soil to improve the stability relating to potential weak zones in the
foundation. The berm material was probably not well compacted.
Since monitoring began in 1962, the dam has experienced relatively large deformations
(vertically up to 850 mm, laterally 750 mm). Figure 11.10 shows typical movement vec-
tors plotted on a time basis. It will be noted that the overall effect is one of spreading and
that the displacement on the crest was downstream until about 1978, thence it has been
upstream.
Figure 11.26 shows the observed settlements on the dam crest. The lateral movements
were also measured. It will be noted that there is a general trend for on-going deformation
which is reducing with time, but there have been rapid movements in 1968, 1983 and
1995. These relate to low reservoir levels and appear to coincide with rainfall events at the
end of the reservoir drawdowns. The movements increased with each drawdown except
that in 1998 the movement was less. This is probably because of a rigorous procedure of
filling cracks in the dam crest as they formed, stopping rainfall infiltration.
In 1995 there was severe cracking and disruption of the dam crest and longitudinal
cracks were observed in the other drawdown periods. Extensive investigation and analy-
sis carried out by Snowy Mountains Engineering Corporation and Woodward Clyde
showed that the factor of safety was low for the upstream slope above the first berm.
However to model the observed behaviour of displacements at the surface and in an incli-
nometer which showed the development of a shear surface at a depth of about 11 m, it
was necessary to model the cracks and softened zones which had formed in the core.
These persisted at least down to full supply level. It was also necessary to model transient
pore pressures which occur in the cracks/softened zones during rain events and pore pres-
sures in the filters. Conventional effective stress analysis gave low factors of safety, but
only for shallow surfaces through the rockfill, i.e. not modelling the real situations. The
observed deformations could only be replicated by numerical modelling if very low mod-
uli were used in the rockfill (in the order of 5 MPa) and weakening of the core.
During construction of remedial works, which included a rockfill berm on the upstream
side and new filters and rockfill for the upper 10 m of the downstream side, a wedge of the
core was displaced nearly 50 mm upstream. This displacement also sheared an inclinometer
Search WWH ::




Custom Search