See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319477792

Settlement behaviour of a major dam

Conference Paper · November 2006

CITATIONS READS

5 1,692

5 authors, including:

John C. Small Harry G. Poulos

The University of Sydney The University of Sydney
182 PUBLICATIONS   2,842 CITATIONS    300 PUBLICATIONS   10,918 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Underpinning of Foundations by Piles View project

Earth and Rockfill Dams View project

All content following this page was uploaded by John C. Small on 05 September 2017.

The user has requested enhancement of the downloaded file.

 SETTLEMENT BEHAVIOUR OF A MAJOR DAM

Peter J Burgessr, Delfa Sarabia', John Small3, H. G. Poulosa and Jayanta Sinhas

ABSTRACT

The assessment of settlement behaviour of clay core rock fill dams has always been a challenge for dam
designers and geotechnical engineers. The method of construction and the material properties of the clay
and rock fill materials used in the dam construction have a significant influence on the inter-zonal
interaction and the load transfer that occurs within the dam. At times this load transfer can lead to
excessive differential and total settlements. The paper presents a case study of a major dam that experienced
large settlements during and after construction. An elaborate analysis has been carried out by modelling
the sequences of construction by using afinite element program (PLAXIS).

The paper describes the influence of the degree of compaction and moisture control on non-linear
deformation characteristics of clay core. High vertical strains in the wet placed region of the core and low
strains in the dry placed regions were analysed for possible shear development between the core and shell.
The rock fill for the dom embankment consists of quartzite, metasiltstone and phyllite material. These
materials have apparently undergone deformation with increasing height of the dam due to softening and
crushing as saturation of the embankment took place. The effect of soil consolidation and strength gains
have been considered in the analysis and are discussed. The settlement behaviour of the dam including these
fficts has been analysed, and compared with the historical post-construction settlements.
This paper is intended to provide valuable informationfor dam engineers handling clay core rockfill dams -
especiallywhere there is excessive settlement of the core.

Of particlllar interest are the movements and

1 INTRODUCTION settlements within the embankment which have
been larger than those experienced on previoLls
Blowering Dam is located on the Tumut River in dams in the area especi ally during construction.
southern NSW approximately 13km south of the However, the rate of settlement has decreased
town of Tumut. The dam was constructed with time, and it is considered that the dam has
between June 1964 and September 1968 to store achieved a state of balance by redistributing
water for irrigation, mainly in the Murrumbidgee forces between various zones. Figr-rre 2 presents a
and Coleambally Irrigation Areas. It is a rockfill comparison of the core settlement of Blowering
dam with a central earth core ll2 metres high Dam with various earth core dams.
with a crest length of 730 metres and contains
6,408,000 cubic metres of rockfill including filters An in-depth study of the settlement of Blowering
and 2,,182,000 cubic metres of earthfill. It has a Dam has been previously undertaken by Hunter
concrete chute spillway located at the right and Fell (2003). This current study should be
abr"rtment (Figure l). seen as a supplementary study to that document
and focuses on the results of an alternative
The design and supervision of construction by modelling approach.
contract were performed by the Snowy Mountains
Authority (SMA) also referred to in this paper as 1 Senior Principal, Coffey Geotechnics
the Snowy Mountains Hydro-Electric Authority 2 Geotechnical Engineer, Coffey Geotechnics
(SMHEA) on behalf of the Water Conservation 3 Senior Principal, Coffey Geotechnics and Professor
and Irrigation Commission of the State of New of Civil Engineering, University of Sydney
South Wales. 4 Senior Principal, Coffey Geotechnics and Emeritus
Professor of Civil Engineering, University of Sydney
5 Project Engineer, State Water Corporation

ANCOLD 2006 Conference Page 1

 Plan of Blowering Darn

\ l$,t*ltt:

iirt,
I*

"-^l l.:' ;r i ,i . , ir ., i ) 't i.: ,r l:

'^""'^'"t "i ,'..., , .'*., r,i.:i ri:,. ..1 .,1
l"'
/
Section atong and Up*tr"earn Elevation nf Blower.ing Oam Spiltway

Figure I Blowering Dam - Plan View - SMHEA

For this study a numerical analysis using finite The dam site is located within a folded sequence
element techniques has been undertaken to model of inter-bedded metamorphic rocks, mainly
the observed core settlement behaviour quartzfte (from sandstone), meta-siltstone (from
incorporating the construction sequences, material siltstone) and phyllite (from shale). This
properties and moisture contents during formation is called the Brandy Mary Beds.
construction combined with the results of
embankment monitoring. The shear strains and Alluvium covered the bedrock in the river channel
stresses have been analysed, and the distribution section to depths of 6 to 15 metres. Clayey
of stresses within the embankment have been slopewash soil with rock fragments similar to
examined during the phases of construction. This material from the earthfill borrow areas covered
paper presents the outcome of the analysis much of the abutments.
highlighting potential explanations of the
settlements. The foundation area under the earthfill and the
filters was stripped of all surface materials above
moderately weathered bedrock and excavation
2 CONSTRUCTION OF THE was continued into bedrock to a non-erodible
DAM foundation.

The cross section and long section profile of the The earthfill zone was constructed using
embankment are shown in Figures 3 and 4. The slopewash soils obtained from borrow areas
nomenclature for the embankment zones adopted within the storage area upstream of the dam.
the standard convention for the SMHEA as These soils were moisture pre-conditioned prior to
follows: uplift, placed in the dam embankment and
compacted with at least 8 passes of sheep-foot
Zone I Earth Core, rollers with a compacted layer thickness of 150
Zone 2 and 2B Selected filter Material, mm to a minimum dry density of 98% Standard
Zone 3 A Inner zorte of selected rockfill, Maximum Dry Density (SMDD).
Zone 3B Outer zone of selected rockfill,
The fine and coarse filters were processed from The procedures for placement of the rockfill
river gravels, wetted and compacted with a called for the Zone 3,A. and the Zone 38 materials
smooth drum vibrating roller in layers to be placed in layers of either 900m thickness or
approximately 0.48 metres thick. 1.8m thickness with water sluicing at a rate of 0.9
to 1 by volume and then compacted by a smooth

t4c€s

'rf,CIOs

E
E
c ts00s
"g
{)
t*
{n
c
*
()
8CI0s

g,
c
L-
:T 6CIos
E
C
{}
E
a 400s
(p
(fl
fll
(}
() ?fffi

o
.50 rGO t$0 *$(l i{5{r

Em,bankment he,ight {rn}

.d iiff :ff'J'ifrT', j 5ffiffr

Figu r e 2 s tru ctio n
:"frffi
The inner zones of rockfill (Zone 34) consisted of drum roller with 4 passes. In places in the
the selected, better class quartzite rock thoroughly downstream of the zones of the rockfill the layer
wetted and compacted by 4 passes of the vibrating thickness was increased from 0.9m to 1.8m and
roller in 0.9 metre layers. The outer zone of still apparently compacted with the same energy.
rockfill (Zone 38) was placed in 1.82 metre layers
and weaker phyllite was permitted.

Section through Blnwerlng Darn

t.s L l{L 38[r.3s m

RIPRAP ^ ^.-
^,

R t. 335,?$ m

$ 1 rrr U'OUoo
RSCKFILL TARTHFILt ftocKFrL!- i
ki,0,t, RL 3SS,3l rr ,,
RL ?$&.7{l r"n

RAhiSCIM FILI.

FSUNPATIfr I'I TXCA VA YI*H LI'{€

**'****-
GfiOUT CURTA iT
BL A r.fi{ tr sRsuTrfrG

Figure 3 Embankment Cross Section - SMHEA

 n tollolbtr or SrU thro (. y g,i/trtrit
tovtll d crtd Q,L. eih*rrfil ,.lkml{ fd;olhlbtr qf .so ?O+15'40
8 Rl.p@ +
g
&t/*%fi
rllttfslotP rrdt}zt rutt rtlolhlro
J dr s?&rJo
5FI
R.L.,oo I ,-,J
.qwEffi{^Tn'
I

f; R L,qx,
+
ea orD r, ; --t3gl:staol
tr R,r.roo t RL.ffiL) t-nL9J6',12
(r*uanra,ba nraa)

11J@ 6100 &tOO tiiOo .,0'OO Jf+OO

Station (100's of feet)
Momrl.r
mcblc rcrh
Long section profile ofBlowering dam (courtesy ofNSW Department ofPublic Works and Scrvices, Dams and Civil Section),

Figu Ye 4 Embankment Long Section Profile (Imperial Units)

The particular characteristics of therockfill and its All seams of soft, friable
and weak material in the
likely behaviour in the dam embankment are an foundation were cleaned out and backfilled with
involved subject beyond the scope of this paper. concrete or covered with pneumatically applied
For a start, the materials used were subject to mortar. The foundation was blanket grouted for at
significant strength reduction upon saturation, least the upstream two thirds of the earthfill
reducing in the case of the phyllite material by as contact. After blanket grouting, a single grout
much as 3 5% to 620/o. curtain was constructed across the foundation.

Furtherrnore the gradings adopted for the rockfill The earthfill was not placed until the foundation
were more suited to a hard rock such as granite treatment and grouting had been completed in the
and gneiss as used elsewhere on the Snowy river section. As a result, the rockfill in the outer
scheme, and had too few fines to give a dense sections of the embankment had been placed to a
compacted mass. It is understood that application height of 17 metres before any earthfill was
of water was intended to induce the strength placed.
reduction that had been noted in the rock strength
testing. After the embankment had risen to a level 40
metres above the foundation, the embankment
In our opinion it is more likely that the sluicing, was constructed in approximately horizontal
although effective in redistributing the fines such layers for the whole width of the embankment.
that the load would be taken by the coarser rock
particles, did not provide the degree of saturation The earthfill core consisted of phyllite and
required to induce particle breakdown at the time quartzite slopewash material from several sources
of placement. Moreover the relatively light within a radius of 3 .2km of the dam site. The
weight roller combined with layer thicknesses up material comprised gravelly clay with excess
to 1.9m would not have been effective in clayey fines to clay soil, of medium to low
compacting the rockfill to a fully compact state. plasticity.

Consequently there was potential for rockfill to All filter materials were processed from naturally
remain as a relatively low modulus mass with the occurring gravel deposits. The rockfill consists of
possibility of creep occurritrg as continued quartzite, meta-siltstone and phyllite.
breakdown occurred.
A moisture range of l% dry of optimum to I%
Reports of the thicker layers on the downstream wet of optimum was adopted for the embankment.
zones suggest that the downstream zones would However, it was considered that the earthfill
be of lower modulus. Accordingly, in setting the should be placed as wet as possible consistent
model parameters, recognition was taken of the with constructability aspects as this would
potential for differing effects upstream and provide a flexible low perrneability earthfill that
downstream and the likelihood of greater would be able to accommodate movements in the
settlement in the rockfill zones as the embankment without cracking.
embankment rose higher.
Initially, in the lower third of the earth core the Period Max. Fill Dry Moisture
water content was slightly dry of optimum, and Height Density Content*
the pore pressures as measured by piezometers (m) (tim3)
were low. Jul 1966- 34-37 r.t9 0.3% drv
Feb 196l
When the earthfill was approximately 35 metres Mar 1967 - 67-74 t.75 0.3% wet
above the foundation, it was decided to aim at a Oct 1967
moisture content slightly above optimum. Nov 1967 - t12 1.78 0.1% wet
Apr 1968
From then on the placement moisture content was Table I : Summarv of Core Moisture Content
kept at an average of approximately 0.3% wet of
optimum. When the embankment had reached a Note: Median values are quoted for the individual
height of about 70 metres, some difficulty was time periods.
experienced in movement of equipment on the
earthfill and the moisture content tatget was *Field Moisture Content Variation from Optimum
decreased to optimum. Moisture Content

Table 1 presents a summ ary of core moisture The total settlement of the core at the end of
content and dry density. construction as measured by the internal vertical
movement (IVM) gauge was about 5.8m. Figure 5
As the eafthfi; was o, u.r, low penneability it presents the settlement of the IVM cross arms.
was assumed that there would be no dissipation of
pore pressures during construction. The During construction, there were two distinct
piezometers located near the axis of the dam, patterns of settlement observed with the change of
which is about the centre of the earthfill, have pattern from one to the other being at
shown no significant dissipation of pore pressure. approximately RL 316m.
There was, however, some dissipation of pore
pressure towards the outer limits. The At about this level, there were two changes in the
piezometers, installed after the placement construction procedures. The placement moisture
moisture content was increased above optimum, content was increased and the earthfill material
showed evidence of pore pressure build up. was obtained from a different borrow area to that
used in the lower parts of the embankment.

lft)ttt

q
// f3

{rrcrl
Figure 5. Core settlement during construction - Imperial Units
 from 8.5m to 9.6m. The finite element model of
Below RL 316m, the earthfill was predominantly the dam is shown in the Figure 6.
quartzite and phyllite slopewash, and above this
level was residual weathered porphyry slopewash, The elastic-plastic Mohr Coulomb model was
which was generally finer material. used for the rockfill, filters and earthfill core. The
rockfill and filters were modelled as drained while
The change in consolidation characteristics the core was treated as undrained. The core being
appeared to be more likely due to different modelled as undrained enables the full
properties of the earthfill materials rather than the development of excess pore pressures.
change in moisture content.
Consolidation analysis was used in all the
construction stages to analyse the development or
3 MATERIAL PROPERTIES dissipation of excess pore pressure with time in
the earthfill core.
Table 2 presents material properties used in the
model. The method of assessment of these
properties is discussed in Section 5. 5 MATERIALPARAMETERS
The material properties of the embankment
4 METHOD OFANALYSIS materi al zones were developed using the available
data including laboratory testing of the materials.
The analysis of the settlement behaviour of
Blowering Dam was carried out using a finite Table 3 presents a summary of the compression
element programPLAXIS. PLAXIS is capable of and recompression indices derived from the
modelling the non-lin ear, time dependent and laboratory consolidation testing carried out on the
anisotropic behaviour of soils and rock. core materials. These indices were used to derive
the modulus of the earthfill core.
An iterative approach was adopted in the
application of the PLAXIS pro gram, whereby the The core was modelled as havine three mam
best assessment of the soil's strength parameters regions:
was made on the basis of the soil test results from . the low er zone placed "dry",
the investigation and design of the dam. o the mid zone placed "wet" and,
Adjustment of layering and loading was then . the upper zone placed at optimum
undertaken to approximate the observed moisture content.
movements.

A staged construction was adopted in the analyses

to simulate the construction sequence of layers.
The dam was initially modelled into three layers
to simulate the materials placed at different
moisture contents. The model was further
d into
elopecl
devel I z Ilayers w th thickn
lnto 12 rcKnesses rangln
Unit Bulk Unit Cohesion Friction Modulus Permeability (cm/sec)
Weight (kPa) angle (MPa)
(kN/m3) (degrees)
Core-dry 20.1 5 27 20-26.8 2 x l0-'
Core-wet 19.7 5 27 r2-t7.9 2 x 10-'
Core-optimum 20.1 5 27 2.5-9.8 2 x 10-'
u/s* d/s*
Zone 2A & 2B 2t.s 5 40 55 45 1x102
Zone 3A'- 1 20.9 5 43 55 45 1 x 102
Zone 3A-2 20.6 5 40 50 40 1x102
Zone 3B 20.6 5 38 40 30 I X 102
Table 2 Summa of Material Pro erties
*u/s upstream zones, d/s - downstream zones
-
 The rockfill was modelled in three zones.
Layer Compression Recompression o Inner Zone 3,A,.- I : finer sized rockfill
Index, Cc Index, Cr placed in 480mm layers. This zone was
I - placed 0.102 0.022 of low compressibility.
drv
2 - placed 0.08s 0.01 I Outer Zone 3A-2: rockfill placed in 0.9m
wet thick layers. This zone is more
3 - placed 0.066 0.01 I compressible than Inner Zone 3.A.- 1

optimum
Table 3: Summary of compression and Zone 38: rockfill placed in 1.82m thick
recompression indices - earthfitl core layers. This zone is of high
compressibility.

The three main regions were further divided into The assessment of the modulus for the rockfill
four layers. The earth core modulus was required a degree of iteration until a modulus was
calculated from the consolidation test results obtained that, in conjunction with the earth fill
together with the stress state of each layer. This modulus gave an embankment deformation
meant that as each layer was added the modulus compatible with that measured from installed
of each layer had to be recalculated and increased instrumentation. One of the determining factors
each time. In this manner the lowest layer has the in the setting of the parameters for the rockfill was
highest modulus. the observed behaviour of downstream movement
at the end of construction condition. To
The filter zone was modelled with low reproduce this behaviour and the subsequent first
compressibility. filling behaviour it was assessed that the
downstream rockfill zones were probably of
slightly different modulus as a result of the effects
of differing saturation conditions upstream and
downstream of the core.

S{i{].{.:'S 51i0.i?{t

,i
i
45,!.*S :
..4

i
1

l
.10rt.0$ I

,.j

':

-l
35$^r}CI

JU1J.Vl.E

,..:

25{].0'0 ,

..:
ltl,[.$'] I :

'.i
l5tl"0s:

Figure 6 -Mesh Adopted for PLAXIS

6 RESULTS OF MODELLING Figure 8 shows the magnitude of strain induced in
the earth core as a result of the settlement. The
The results of the modelling are summarised in pattern of strain induced in the core is one of
the following four figures: centre line symmetry with steeply sloping zones
reaching a magnitude of about 12% in the centre
Figure 7 shows the total maximum deflection at of the core. The shear strains intersect at about
the end of construction. mid point in the earth core coffesponding with the
portion of the core compacted wet of optimum
Figure 8 shows the coffesponding maximum shear and in which about 50% of the core settlement
strains developed within the embankment at the had occurred.
end of construction.
Figures 9 and 10 show the coffesponding results
Figure 9 shows the total maximum deflection at of the modelling for the first filling condition.
the end of first filling. This was modelled using the load of the water on
the upstream face to induce the observed
Figure 10 shows the coffesponding maximum settlement of the core. The resultant settlement of
shear strains developed within the embankment at 6.17m compared with 6.3m as measured by the
the end of first filling. IVM gauge.

7 DISCUSSION 8 CONCLUSIONS

Figure 7 represents the base case for the dam The modelling carried out during the current study
embankment. For this case it was found that the has, in the opinion the authors provided a valuable
strength and stiffness of the enclosittg rockfill interpretation of the deformations that can occur
zones had a significant effect on the settlement of in a large earth and rockfill dam. The
the clay core. The modulus of the rock fill was deformations and highest strains occurred in the
adjusted until the direction and magnitude of the middle two thirds of the dam embankment core
movement were of the same order of magnitude as which happens to coincide where the designers
the results of the embankment monitoring. The sought to place earthfill materials with the greatest
resultant settlement of 5.76m compares with 5.8m water content with respect to optimum moisture.
as measured by the IVM gauge.

klll
:.1..it.",1

{.J:r&l

' ].'fi*$

i,2rx

2.BrX

i*;') fir"lll
i:.lii::i:l

i:r,i*
hEst *,?vk'

ffi 1,2(.tll

u.Bfiil
m
tilRiJ
l*i
- -...
{J..t(s.J
i1

Figure 7 Total maximum deflection 5.76m - end of construction w;,"

 i:$l

r5l.t6lil
I
r'*'""''
I
-,,...-.."
N-'
N
Nhi,i$r)

, tlfr.i{til

' -32.t]*(r

,i$.sl)C

i;4'its;)0

, , 'i2t3,{"i*L'}

:t

N16.{ii"l.t
N
sHl2.s(uj
w
ffi

ffi R,';'**

ryq't"(i$ri
::
Figure 8 Extreme Shear Strain 12oh- end of construction
H 'J"($ii

E
n..1.*[ri]

l,rr l

Figure 9 Total maximum deflection 6.17m - end of first filling

Figure 10 Shear Strain l2o/o end of first filling

In addition, the dam was constructed using a rock
fill material that was particularly susceptible to
loss of strength upon saturation. The rock fill
materials appear to have been compacted in a
manner more suited to a hard rock such as fresh
granite or gneiss and with a layer thickness gre ater
than can be effectively compacted where
maximum particle breakdown during compaction
is required.

For these reasons we consider that the settlement

characteristics observed for the dam embankment
can be related to a combination of moisture
content variation within the central core zone of
the dam combined with a lack of substantial
stiffness from the enclosing rock filI zones.

The maximum shear strains of about 12% have

developed in the central part of the core where the
placement moisture content is highest relative to
optimum and are considered unlikely to have
formed continuous failure planes.

ACKNOWLEDGEMENT
The authors would like to thank Mr Bob Barber
and Mr Glenn Tully, State Water Corporation,
New South Wales for their review and permission
to publish some of the analysis results in this
paper.

10 REFERENCES
Hunter, G. & Fell, R. 2003. The Deformation
Behaviour of Embankment Dam. UNICIV Report
R-416.

Fell, R. MacGregor, P. Stapledon, D. & Bell, G.

2005. Geotechnical Engineering of Dams.

Snowy Mountains Hydro-Electric Authority, 3rd

Edition 1993. Engineering Features of the Snowy
Mountains Scheme.

State Water, 2005. Blowering Dam Surveillance

Report July 2005.

View publication stats