You are on page 1of 5

CONCRETE DAMS ON ROCK FOUNDATIONS

INTRODUCTION where a uniform foundation rock foundation exists the pressure head
A detailed analysis of the control of seepage under concrete dams drops to a value close to the tailwater head at the line of drainage
has been presented in the First Rankine Lecture by Casagrande (1961) control wells and then remains constant.
from which these notes have been precised and all the figures are
taken. From a geotechnical point of view, once the competence of the ANALYSIS OF DRAINAGE
rock has been established the seepage below the foundations and Figure 2 shows the theoretical results for various locations of an
around the abutments is the most important factor requiring control. interface (rock to dam) horizontal drain upon the foundation uplift
The seepage loss is generally of secondary importance and in fact will pressures. The rock is assumed to be semi-infinite in extent, isotropic
generally be increased by the seepage controls. and the water flow to obey Darcy's law. The closer the drain is to the
heel the greater the reduction in uplift pressure, however the shape of
CONTROL OF SEEPAGE THROUGH ROCK the uplift pressures is considerable different to that normally assumed.
Control of seepage is achieved, in general, by both a grout curtain This is not surprising since a surface drain is hardly a good
under the dam 1.2 to 3 metres from its heel followed about 6 metres approximation to a line of deep penetrating wells.
further from the heel by a line of drainage wells. Single line grout
curtains are typically formed by grouting into 75 mm diameter
vertically drilled holes spaced at 1.5 metre centre to centre and 12 metre
deep with every third hole extended to 24 metres depth. Drainage wells
are similarly sized and spaced with a depth of 12 metres. A typical
example is shown in Figure 1 (Casagrande, 1961).

Figure 2. Uplift pressures resulting from surface line drainage.

The designer should be aware that the surface layers of rock are
likely to be more permeable than the deeper layers and that the
hydrostatic pressures against the dam face are likely to open up rock
joints along the heel of dam. This may cause the full hydrostatic
pressure to act through the full depth of the pervious zone of the rock
(at the heel). If hydrostatic pressures exist below the heel throughout
the relatively pervious rock then the conditions shown in Figure 3 are
approximated. From this it is evident that the seepage control wells
should extend through the full depth of the pervious rock zone.

Figure 1. Foundation uplift pressures, Hiawassee Dam.


Figure 3. Examples of uplift control from fully penetrating wells.
The seepage control, using these measures, is assumed to produce Note: d = 50 ft; hw = well level below tailwater;
uplift pressures that start as equal to the reservoir head at the heel, drop hm = water level mid-point between well and toe.
linearly to 25 to 33% of the hydraulic head across the dam at the line of If the conditions in Figure 3 are approximated then the theoretical effect
drainage control wells and then drop linearly to the tailwater head. In of the line of seepage control wells may be assessed by solutions stated by
practice, and in agreement with theory developed by Brahtz (1936), Musket (1937). These solutions are stated and illustrated in Figure 4 with

CIVL-842 Notes: Concrete Dams on Rock Foundations - Prof. G.P. Raymond. 1


drawn elevations for the case of a = d. In practice a < d resulting in lower
required values of well depressed water head hw. The figure shows a
pressure head across a vertical entrance face causing water to flow into a
confined pervious strata of thickness D. An infinite row of wells are
shown parallel to and at a distance d from the entrance face. These wells
penetrate fully the strata and have a of radius rw and are spaced, centre to
centre, at a distance a. The hydraulic head is hc (difference between
upstream and downstream water head). For any given values of hc, d, a,
rw there is a calculable depressed well head (below the downstream head)
of hw that ensures all the seepage entering the strata is removed by the
wells. The head mid-way between the wells is given as hm above the well
water elevation. The water elevation beyond the wells is equal to the
downstream water elevation (approximately). Some numerical values are
given in Figure 3 with d = 5a and d = 10a. The seepage is simply
calculated as the flow under an hydraulic gradient hc/d. The solutions for
pressure relief show that from a practical view point a line of closely
spaced small diameter wells are more efficient than a line of larger wells
spaced at greater spacings.

Figure 6. Examples for well water at tailwater elevation.

Figure 4. Theoretical uplift control from line drainage wells. Figure 7. Theoretical uplift for well water above tailwater.
Figure 5 gives theoretical solutions, developed with some simplifications ANALYSIS OF IMPERFECT CUT-OFFS
and the limitations of a/d < 3 and rw/a < 0.1, to the case of the drain water As mentioned earlier it is common to use a grout curtain in addition
level higher than the tailwater level. Two numerical solutions are given to the line of drain wells. Dachler (1936) presents a theory for an
in Figure 6. These solutions show that for practical spacing of the drain imperfect cut-off and Figure 8 gives solutions for this theory. The theory
wells (75 mm diameter holes on 1.5 or 3 metre spacing) and a well drain assumes that there are n equally spaced slits. Whether the imperfect slits
water level (elevation) equal to the tailwater level (elevation) the uplift are vertical or horizontal is of no theoretical significance. D is the total
pressures are only slightly increased (4.5% and 6.8% respectively). strata thickness, d is the (sum of) the thickness of the impermeable portion
of the strata giving W = D - d as the total thickness of the open space.
Then d/D = cut-off ratio, W/D = open space ratio, and Ec = the cut-off
efficiency and is defined as (q - qc)/q where q is the rate of flow without
a cut-off and qc is the rate of flow with a cut-off. Dachler's solution can
be simplified for W/D < 0.1 and the solution for B/D = 1 and two values
of n (n = 20 and 120) are shown in Figure 8. If n is large (n > 20) then
the efficiency, Ec decreases approximately inversely proportional with n
(the number of openings). (Note: This finding loses its validity when
dealing with thick walls).

Figure 5. Theoretical uplift head uw at well above tailwater level.

Figure 7 gives the solution for the case of a well drain water elevation
controlled at a higher elevation than the tailwater elevation.

CIVL-842 Notes: Concrete Dams on Rock Foundations - Prof. G.P. Raymond. 2


with drain well water elevation above tailwater elevation: This results
in an uplift diagram which shows a linear pressure drop to approximately
the drain well water elevation and then linear to the tailwater elevation.
Patterns (c), (d), and (e) assume there are no unusual geological
complications and that the drain holes extend completely through the
pervious rock stratum. The shallow drain holes previously used on most
dams have fortunately been quite effective, indicating that, at the majority
of sites, the pervious foundation zone is relatively shallow. This may not
be true for abutments (the ends of the dams) for which unfortunately very
few observations are available.
Pattern (f), (g), and (h) are three examples of the deviations from the
ideal pattern that are sometimes encountered.
(f) Partial penetration of drain holes: This should cause a large bulge
in uplift pressures downstream of the drains, similar to the Brahtz (1936)
Figure 8. Theoretical efficiency of imperfect cut-off curtain. solution illustrated in Figure 2.
THEORETICAL EFFICIENCY OF A SINGLE-LINE GROUT (g) Stratified rock dipping upstream, having a pervious layer
CURTAIN sandwiched between the rock and a deeper impervious strata and cut-
There are so many variables that a definitive result of the theoretical off by the dam toe, a single-line grout curtain extending to the
efficiency of a single-line grout curtain is of little value. On the other impervious layer, while the drain holes stop short of the impervious
hand even with a small open space ratio of 0.1% and a value of n = 20 the layer: Such a combination could cause an uplift diagram that shows a
efficiency is only 29% (i.e. 71% of the flow goes through the slits). For downstream zone beneath the dam where the uplift pressures are almost
the same open space ratio this efficiency drops as n increases. In equal to reservoir uplift pressure. This dangerous condition would only
conclusion it is dangerous to expect a single-line grout curtain to be highly be noted if uplift measurements were being observed during reservoir
efficient. Either provide an upstream impervious blanket or move the line filling. In such a case the reservoir water level could be lowered, the
of drain wells further downstream. cause investigate, and the drain holes extended through the pervious layer
with possible supplementary drainage holes near the downstream toe (and
HYPOTHETICAL EXAMPLES through the dam) into the pervious stratum.
(h) assumes a similar geological situation to (g) except that the
pervious layer does not come to the surface, and that downstream of
the dam it is cut off by an impervious fault: The drain holes will
control the uplift along the base of the dam at approximately tailwater,
particularly if a slightly more pervious surface zone should exist. This
will create a false sense of security since dangerously high uplift pressures
in the pervious stratum not far below the base of the dam, and which
extend downstream of the dam, will exist. Drilling the drain holes much
deeper, or along the downstream toe of the dam, could possible eliminate
all danger. To discover such conditions during filling both base uplift
measurements and numerous piezometers extending to various depths into
the rock would be required. A similar actual example was discussed by
Terzaghi (1929).

CONCLUSIONS FROM EXAMPLES


(1) Since drain holes are cheap relative to total costs they should be
Figure 9. Hypothetical examples. drilled to a depth at least half the height of the dam rather than the 12 to
15 metres used for most concrete dams in the past. Local geologic
Some hypothetical uplift examples are considered in Figure 9 for conditions may indicate the need for still greater depths.
different assumptions concerning the grout curtain, the line of drains, and (2) Where any doubt exists of the geologic conditions that might alter
the geologic conditions. the predicted hydro-static pressures in the rock zones below the dam
(a) Perfect grout curtain: A perfect cut-off results in a water level equal and/or in the abutments, uplift measurements along the base and
to the full water head upstream of the grout curtain and equal to the piezometer observations at many locations should be recorded.
tailwater elevation downstream of the grout curtain.
(b) No drainage and near perfect grout curtain: the water level would UPLIFT OBSERVATIONS FOR TENNESSEE VALLEY
drop linearly from the full water head at the heel to the grout curtain and AUTHORITY DAMS
then linearly but less rapidly to the toe. A 50% drop in seepage, which The measured uplift diagram for the Hiawassee Dam is shown in
would be considered an effective single-line grout curtain, would Figure 1. It consists of a straight line from reservoir level to tailwater at
generally be unsatisfactory from a cost-benefit basis. the line of drains, and then practically a horizontal line at tailwater
(c) A near perfect grout curtain and a line of complete penetration level for the area downstream of the drains. It is an ideal example
drain holes: This will result in an uplift diagram which still shows a rapid confirming the preceding theoretical analysis for the performance of
linear pressure drop to the grout curtain followed by a less rapid drop to drainage wells. The foundation rocks at this dam are steeply dipping,
the drain well water elevation. intensely jointed quartzites, and schists.
(d) No grout curtain and a line of complete penetration drain holes Figure 10 shows the average uplift measurements for four T.V.A.
with drain well water elevation slightly below tailwater elevation: This dams (Fontana 146 m, Hiawassee 94 m, Cherokee 62 m, and Douglas 53
results in an uplift diagram which shows a linear pressure drop to m). Upstream of the drains the water level drops from reservoir level in
approximately the drain well water elevation. the form of a slightly concave curve, with no indication of the location of
(e) No grout curtain and a line of complete penetration drain holes the grout curtains. The average water level in the drains is

CIVL-842 Notes: Concrete Dams on Rock Foundations - Prof. G.P. Raymond. 3


approximately at tailwater level and downstream from the drains about
10% of a linear theoretical drop for the condition without drains. While
this is very safe it indicates that the drain holes would have given better
control had they been deeper.

Figure 12. Data from eight Bureau of Reclamation Dams.

Figure 10. Data from four Tennesse Valley Authority dams.

Other examples, shown in Figures 11, 12 13 and 14, and four case
examples from Terzaghi's private files are then describe to suggest that
grouting was not as effective as general thought. To quote Terzaghi
"many owners are penny-wise and pound-foolish by spending money on
grouting and then little or nothing on instrumentation to establish the
effectiveness of the grout curtain".

Figure 13. Data for Grande Coulee and Shasta Dams, Bureau of
Reclamation.

SEEPAGE THROUGH ABUTMENTS OF THIN-ARCH DAMS


The control of seepage through thin-arch dams is much more difficult
than gravity type concrete dams with a wide base. Thin-arch dams can not
accommodate both a reliable grout curtain consisting of three lines of
grout holes and a drainage system 6 metres from the grout curtain. The
stability of the rock downstream of a thin-arch dam is thus critically
important. Any grouting imposes great difficulties and may do more harm
than good, even if carried out after the dam is built. Indeed there may be
no satisfactory way of grouting in steep abutments for thin-arch dams, in
Figure 11. Uplift for Fontana Dam, Tennesse Valley Authority.
which case the drainage becomes the only line of defence. A major
problem occurs in cold climates from freezing of the outlet water.
Widening of the thin-arch would go a long way to allowing an internal
drainage tunnel. Where the geological conditions can not ensure safe
control of water thin-arch dams should not be considered for that site.

CIVL-842 Notes: Concrete Dams on Rock Foundations - Prof. G.P. Raymond. 4


Terzaghi, K., (1929). "Effect of minor geological details on the
safety of dams", American Institute of Mining and Metallurgical
Engineers, Technical Publication 215, p. 30.

SUMMARY POINTS OF TEXT


1. Seepage is a major concern and the most important factor requiring
control in the construction of concrete dams on competent rock
foundations.
2. Single line grout curtains are not an effective measure for controlling
seepage and uplift pressures under dams and should not be relied
upon for this purpose since even extremely small gaps in the
grouting result in large flows passing through.
3. Drainage wells provide an economical and reliable means to control
seepage through rock foundations and reduce uplift pressures below
Figure 14. Data for Hoover Dam, Bureau of Reclamation. concrete gravity and arch-gravity dams.
(a) before extensive additional drainage and grouting, 4. Locating the drain close to the heel gives the largest reduction in
(b) after extensive additional drainage and grouting. uplift pressures.
5. Small diameter, closely spaced, deep (at least half the height of the
CONCLUSIONS AND RECOMMENDATIONS dam - penetrating the pervious zone) drainage holes are the most
(1) Seepage through rock foundations and abutments may be effective and efficient.
controlled reliably and economically for most dams by drainage measures. 6. Caution should be taken with thin-arch dams since they provide
Even a shallow line of drain holes is effective for controlling uplift special difficulties for seepage control and have small contact areas
pressures below most concrete gravity and arch-gravity dams. This with their abutments and foundations for seepage and drainage
control could be improved, even for very unfavourable geologic control measures.
conditions, by using deep drain holes. The effects are approximately
predictable by the of formulas presented.
(2) Most single-line grout curtains in rock foundations and abutments
of dams have been relatively ineffective in reducing seepage losses. Their
effects should not be relied on in stability analysis. There is an urgent need
for comprehensive observations on the effectiveness of such grout
curtains.
(3) There is a need for more reliable methods to determine in advance
whether and where grouting is needed. Present geological investigations
and " water tests" in drill holes may be misleading. For concrete gravity
and arch-gravity dams the upstream third of the foundation area to a depth
greater than one-half the height of the dam should be investigated.
Excessively pervious zones may require three rows of closely spaced
grout holes to create the necessary width of grouted rock mass. All
exploratory holes should be considered for use as observation wells, as
drainage holes, or as grout holes.
(4) On some projects methods other than grout curtains could be
more effective and economical at reducing seepage. For example (a) an
impervious earth fill against the lower portion of the upstream face of the
dam, (b) sloping the lower part of the upstream face of the dam to prevent
opening up of a crack between earth and concrete, (c) for operating dams
a upstream blanket just upstream of the dam, (d) move the line of
drainage holes further downstream (it is customary to locate the drains
approximately 10% of the base width from the upstream face of the dam,
this distance could readily be doubled), (e) when the drain holes positively
control uplift the drains could be move further downstream.
(5) In contrast to gravity and arch-gravity dams, it is extremely
difficult to control seepage for a thin-arch dam. Special difficulties result
from blasting and grouting operations. These can easily cause more harm
than good. Even a comprehensive drainage system may provide
insufficient protection against development of dangerous hydrostatic
pressures in the rock downstream of a thin-arch dam.

REFERENCES
Brahtz, J.H.A., (1936). "Pressures due to percolating water and their
influence upon stresses in hydraulic structures" Transactions 2nd
Congress Large Dams, Washington, 5:43-71
Casagrande, A., (1961). "Control of seepage through foundations
and abutments of dams", Geotechnique, 11:161-182.
Dachler, R., (1936). "Grundwasserstromung", Springer, Vienna,
p. 82. ("The flow of water in the ground").

CIVL-842 Notes: Concrete Dams on Rock Foundations - Prof. G.P. Raymond. 5

You might also like