You are on page 1of 10

ROCK MASS FOUNDATION SEISMIC TOMOGRAPHY AND ITS

CONTRIBUTION FOR THE MODELLING OF BAIXO SABOR


DAM’S STRUCTURAL BEHAVIOUR

Rogério Mota *, Maria J. Coelho*, J. Piteira Gomes *, A. L. Batista*, Jorge Neves†,


Celso Lima †, D. Silva Matos †
*
Laboratório Nacional de Engenharia Civil
Av. do Brasil, 101, Lisboa, Portugal
e-mail: rmota@lnec.pt, webpage: www.lnec.pt

Keywords: Baixo Sabor dam; Crosshole seismic tomography; Rock mass foundation;
Structural behaviour

Abstract. Crosshole seismic tomography characterization of Baixo Sabor dam’s rock


mass foundation after the first filling of the reservoir is presented. A continuous
foundation profile along the dam’s reference line was obtained with P-wave seismic
velocity distribution. This zoning was used to adjust the foundation mechanical
parameters of dam´s finite element model. The model, developed by LNEC, included a 3D
representation of the dam and its rock mass foundation, was used for the analysis and
interpretation of dam´s structural behaviour during the first filling of the reservoir and
the subsequent operation period.

1 INTRODUCTION
Baixo Sabor dam, located on Sabor river (NE Portugal) is an EDP engineered and
commissioned dam.
The arch dam is a 123 m high structure, with a total crest length of 505 m and a total
concrete volume of 670 000 m3. For the full storage level located at elevation (234) the
reservoir capacity is of 1 095 million cubic meters. A controlled surface spillway is located at
the central part of the dam crest with a discharge capacity of 5 000 m3/s, including four spans
controlled by radial gates and provided of a downstream plunge pool.
In geological terms, the foundation of the Baixo Sabor dam lies on a granitic rock mass.
Crosshole seismic tomography is a geophysical technique that can provide a high -
-resolution imaging of the subsurface, in terms of seismic P-wave velocity (VP).
Consequently, this is an especially adequate technique to characterize, zone and delineate
weak areas on rock mass foundations of large structures, as demonstrated in several
studies1-9.
Seismic wave velocities are related with the quality and strength of the rock masses
(mechanical properties) and so they can be used to evaluate the effectiveness of
foundations treatment, if time-lapse seismic tomographies are carried out before and after
the treatment4,5. Generally, the rock mass consolidation treatment consists in grout


EDP Produção SA, Direção de Engenharia de Barragens

1
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

injections to fill joints that usually are open or filled with soft erodible materials
(previously removed before the injection).
The use of 32 vertical boreholes drilled in Baixo-Sabor dam’s foundation, specifically
for seismic crosshole tests, allowed to obtain a VP crosshole seismic tomography profile
of the entire dam’s foundation along its reference line. The seismic tests and
corresponding tomographies were carried out before (Phase 1)10 and after (Phase 2)11 the
dam’s foundation global treatment (consolidation and grout curtain), and after the first
reservoir filling (Phase 3)7.
Seismic tomographies evidenced low and high velocity zones in the rock mass,
agreeing, in general, with the previous geological-geotechnical zoning of the dam’s
foundation. Phase 2 time-lapse seismic tomographies, and consequent analysis of
variation of seismic velocity in the rock mass between Phase 1 and Phase 2, indirectly
allowed assessing the foundation treatment effectiveness. The present work deals with the
contribution of Phase 3 data to the foundation zoning of the finite element model , which
supported the analysis and interpretation of dam´s structural behaviour during the first
filling of the reservoir.

2 GEOLOGICAL AND GEOTECHNICAL SETTING


The Sabor river stretch upstream of the Baixo Sabor dam has a S42 oW direction and
inflects to S25 oW at the dam site, following the trace of the NE-SW river bed geological
fault zone. The valley is deep, slightly asymmetric, with an open V-shaped form and a 25-
40m wide thalweg. The bank inclination varies between 20º at higher topographic levels
and 40º at lower levels.
The geological and geotechnical investigation works12 performed in the design phase
included the detailed surface geological mapping of 13 trenches and 6 exploration
galleries, geophysical site investigations, 38 rotary boreholes and several in situ and
laboratory tests13. The data obtained during these exploration and characterization works
and the one collected during construction phase excavations detailed mapping, and the
extensive drilling for the foundation treatment, allowed a detailed knowledge of the
foundation rock mass.
Baixo Sabor dam is founded in a granitic rock mass that intruded the turbiditic filite -
greywacke metassediments of the Douro-Beiras Group during phase 3 of the Hercynian
orogeny, approximately 300 m.yr. ago. From a petrographic point of view, this rock
corresponds to a medium to coarse grained, biotitic-muscovitic, porphyroid granite.
This region main tectonic feature is the Vilariça fault, located 6.5 km to the west of the
dam site. It is a late Hercynian NNE-SSW sinistral strike slip fault, with an accumulated
displacement of about 7.5 km and a length of more than 200 km, extending from Sanábria
region (Spain) to Serra da Estrela region, in the center of Portugal.
The rock mass geomechanical zones characteristics presented in Figure 1 correspond to
the situation prior to the grouting treatment and allow to conclude that, in general, the dam is
founded in a fair (GZ2) to good (GZ1) quality granite, locally intersected by NNE-SSW to
NE-SE geological faults filled with gouge, generally less than 0.5 m thick with associated
zones of fractured and weathered rock (GZ3) with variable thickness. At the right bank, two
NW-SE faults intersect the dam foundation.
The fault at the left abutment and the one at block 28-29 (Figure 1) have gouge filling with
thickness of 4 m and 1 m, respectively. The valley fault zone is limited by two subparallel
NE-SW faults, filled with gouge less than 0.5 m thick, and has a heterogenous rock mass
(GZ2 and GZ3) intersected by secondary faults.

2
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

Figure 1: Dam foundation geological map and geomechanical zoning (adapted and simplified from
GEOAREA, 2015 14) and geomechanical parameter range values (adapted from EDP, 2005 12).

Figure 2: Dam foundation geological section and geomechanical zoning (view to downstream).

The rock mass is also intersected by several pegmatite and quartz veins, generally less
than 0.5 m thick, that intruded pre-existing fractures, except for the large quartz vein
associated to the fault at the left abutment that, locally, can have a thickness of more than
20 m.
The interpretative geological section presented in Figure 2 depicts the main geological
features of the rock mass and the distribution of geomechanical zones, which is controlled by

3
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

the presence of nearby geological faults, veins and joint pattern, as well as by the superficial
weathering at upper levels of the rock mass.

3 CROSSHOLE SEISMIC TOMOGRAPHY

3.1 Data acquisition and methodology


Crosshole seismic tomography is a widspread geophysical method to evaluate the
elastic properties of materials in the crosshole section, for a variety of applications,
including engineering site characterisation and foundation evaluation2,6,8.
In the crosshole seismic tomography here considered (Phase 3), P-wave velocity (V P)
distribution is reconstructed from a multitude of travel time ray paths, corresponding to
several crosshole seismic measures as illustrated in Figure 3, assuming a purely
bidimensional wave propagation on the crosshole section. So, for each seismic source
activation along the source borehole, there is a multi-channel seismic record from the
receivers array along the receiver borehole.

Figure 3: Data acquisition layout for crosshole seismic tomography

All P-wave measured travel times are jointly inverted into a V P matrix assigned to a
cells grid which discretizes the area with seismic ray paths. The inversion is performed by
a SIRT (Simultaneous Iterative Reconstruction Technique) type algorithm6,9, that assumes
straight ray paths (direct waves), constant velocity for each grid cell and isotropic
materials. It starts with an initial velocity matrix for which both the travel times for all the
straight ray paths and the differences between the theoretical and the measured travel
times are computed. These time differences, also called residual times, are the key to
improve the velocity model in the consecutive iterations until the desired accuracy is
achieved.
In Baixo-Sabor dam’s foundation, 31 crosshole seismic tomography tests were carried
out using 32 vertical boreholes with lengths between 46 and 52 m, specifically drilled
along the drainage gallery, for the seismic tests execution (see boreholes implantation on
Figure 1).
Seismic data were acquired in three different times: the first one or Phase 1 10, before
any foundation treatment, the second one or Phase 2 11, after the global foundation
treatment, and the third time, or Phase 3 7, after the first reservoir filling – being the last
one the object of the present work.
Test boreholes were steel cased, sealed at the bottom and filled with water, in order to
carry out the seismic crosshole tests.
In Phase 3 electrical detonation caps were used as seismic sources each meter, and as
receivers, two multi-channel cables, each having 24 hydrophones moulded in with 1 m

4
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

spacing. This layout conjugated with the high borehole length generated a huge amount of
data gathered – more than 60 000 P-wave travel times.

3.2 Results
Due to the high amount of seismic data it was not possible to jointly process and invert all
the 31 sections for a unique tomographic planned profile. To overcome this setback five
contiguous tomographic planned profiles were computed and then interpolated with Surfer®
software, to produce one single profile (Figure 4). Each one of the five planned profiles was
discretized with 1.5 m side square cells grid used for the tomographic inversion.
In the colour scale used for the tomographic image (Figure 4) the red brown colour was
attributed to the lowest velocity values in order to highlight the massif zones which probably
have relative lower mechanical quality than the remain rock mass. The lowest velocity values
occurred in the left bank first section, in the bottom valley, and in the right bank.

Figure 4: Baixo-Sabor dam’s foundation interpolated seismic tomography for Phase 37 (vertical red colour dash
traces are the limits between the five tomographic planned profiles), view to downstream

The global tomographic profile has identical features, mainly the relative low velocity
zones, from that obtained on Phase 211. It also has a good fit with the geological-
geotechnical zoning (Figure 1 and Figure 2) of the dam’s foundation. The relative low
seismic velocity zones are well correlated with the weathered rock zones (GZ3), and also
with the borehole logs.
This agreement validates the use of the VP tomographic zoning as a contribution to the
finite element model that supports the analysis and interpretation of dam´s structural
behaviour, and for the calibration of the foundation mechanical parameters to be used in
the model.

4 ANALYSIS AND INTERPRETATION OF THE DAM-FOUNDATION


STRUCTURAL BEHAVIOUR

4.1 Structural model


The dam-foundation behaviour during the first filling of the reservoir was analyzed and
interpreted on the basis of the monitoring results and of a continuous structural model
(because no significant movement variations were observed in the contraction joints during
the first filling)15.
The structure of the dam was approximated by a set of supposedly continuous and
homogeneous blocks, considering 30 monthly construction stages, with a rheologic behaviour

5
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

of the concrete characterized by expression (1) and by Poisson’s ratio νc=0.2.


The rock mass foundation was divided in 7 zones of different deformability, as a function
of the results obtained from both the described geomechanical characterization and the
seismic tomography obtained along the drainage gallery after the first filling of the reservoir.
The Poisson’s ratio was assumed with a value of νf=0.2 for all these 7 zones. No time effects
on the foundation behaviour have been considered.
For the dominant concrete type used in the dam construction, the following BaP creep law
was adjusted, considering the results of a set of in situ and laboratory tests,
J(t, t 0 ) 
1
40.0

1  3.2 (t 00.35  0.05)( t  t 0 ) 0.12  (GPa 1 ) (1)
The finite element mesh (Figure 5) has a total of 14 854 nodal points for 2795
isoparametric cubic 20 nodes finite elements (1067 elements of the dam’s body and 1728
elements of the rock mass foundation). The dam’s body has 3 elements across the thickness.
The spans and the shapes of the surface spillway were explicitly represented in the finite
element mesh.

Dam’s concrete (month of casting): Rock mass foundation (Young modulus):


01-2012 01-2013 Ef1 = 20 GPa
02-2012 02-2013
03-2012 03-2013 Ef2 = 30 GPa
04-2012 04-2013
05-2011 05-2012 05-2013
Ef3 = 25 GPa
06-2011 06-2012 06-2013
Ef4 = 20 GPa
07-2011 07-2012 07-2013
08-2011 08-2012 08-2013 Ef5 = 15 GPa
09-2011 09-2012 09-2013
10-2011 10-2012 10-2013 Ef6 = 10 GPa
11-2011 11-2012
12-2011 12-2012
Ef7 = 40 GPa

Figure 5: Dam-foundation finite element mesh considered in structural analysis (view to downstream)

4.2 Actions variation during the first filling of the reservoir


The first filling of the reservoir lasted about 28 months (it began in December 2013 and
was completed in April 2016, when the water level reached the elevation of 234 m).
As the uplifts had small variations during the first filling, the water loads were represented

6
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

only by hydrostatic pressure increments on the upstream surface of the dam, considering
γw=10 kN/m3. The thermal variations in the dam’s body were computed based on the
monitored temperatures in concrete, air and water, considering a numerical procedure to
spread the observed values in a set of discrete points to all the mesh nodes of the dam’s
structure. A linear thermal dilation coefficient was considered for the concrete α=1.1x10-5/ºC.
The analysis comprises the first filling period and a subsequent operation period until late
January 2017. Actions due to water pressure and thermal variations were discretized
fortnightly.

4.3 Some relevant results in the foundation


Figure 6 shows the set of 16 rod extensometers installed in the drainage gallery to measure
vertical and horizontal displacements at the dam insertion surface. Six of which, the three
closest to each bank, are single rod and the remaining ten, in the central part of the dam, are
double rod extensometers.

LB J2 J3 J4 J5 J6 J7 J8 J9 J10 J11 J12 J13 J14 J15 J20 J21 J22 J23 J24 J25 J26 J27 J28 J29 J30 J31 J32 RB
J16 J17 J18 J19

EF2 EF14

EF1 EF15

EF4.1
EF12.1
EF3 EF13

Insertion
Surface EF5.1 EF6.1 EF11.1
EF10.1
EF9.1
EF7.1
EF8.1
EF4.2 EF12.2

EF5.2 EF6.2 EF10.2 EF11.2

EF9.2

EF8.2

EF7.2

Figure 6: Rod extensometers installed in the drainage gallery to measure the displacements at the dam’s
insertion surface.

Figures 7, 8 and 9 show the displacements observed over time (from July 2014 to January
2017) on three rod extensometers (EF2, located on the bottom of the left bank block 7-8 at
177.8 m elevation; EF9, located downstream on the bottom of the central block 17-18 at
116.3 m elevation; EF11, located downstream on the bottom of the right bank block 20-21 at
129.25 m elevation) and the correspondent computed displacements, considering the effects
of water pressure, thermal variations and creep. The influence of the water level and of the
thermal variations on the structural response should be pointed out. The adjustment achieved
between monitoring and calculated displacements can be considered globally good.

7
Extensómetro de fundação EF2 - Horizontal
Rogério Mota, Maria J. Coelho, J. PiteiraBloco
Gomes,7-8
A. -L.Cota
Batista,
177,8Jorge
m Neves, Celso Lima, D. Silva Matos
(mm) (m)
5 240

4 230
to the valley

3 220

Water level in the reservoir


2 210

1 200

0 190

-1 180
to the bank

-2 170
EF2 - Rod with 39.2 m(monitoring)
-3 160
EF2 (computed)
-4 150
Reservoir water level
-5 140
2014-07-01

2014-08-30
2014-09-30

2014-11-30

2015-03-01

2015-05-01
2015-05-31

2015-07-31
2015-08-30

2015-10-30

2016-01-29

2016-03-30
2016-04-30

2016-06-30
2016-07-30

2016-09-29

2016-12-29
2014-07-31

2014-10-30

2014-12-30
2015-01-29

2015-03-31

2015-07-01

2015-09-30

2015-11-30
2015-12-30

2016-02-29

2016-05-30

2016-08-29

2016-10-29
2016-11-29

2017-01-28
Extensómetro de fundação EF9 - Jusante inclinado
Bloco 17-18 - Cota 116,25 m
Figure 7: Displacements measured and computed on the horizontal rod extensometer EF2, located on the bottom
of the left bank block 7-8 at 177.8 m elevation, from July 2014 to January 2017

(mm) (m)
3 240
Upwards

2 230

1 220

Water level in the reservoir


0 210

-1 200

-2 190

-3 180
Downwards

EF9 - Rod 1 with 79.8 m (monitoring)


-4 170
EF9 - Rod 2 with 22.1 m (monitoring)
-5 EF9 - Rod 1 (computed) 160
EF9 - Rod 2 (computed)
-6 150
Reservoir water level
-7 140
2014-07-01

2014-08-30
2014-09-30

2014-11-30
2014-12-30

2015-03-01

2015-05-31

2015-08-30

2015-10-30
2015-11-30

2016-01-29
2016-02-29

2016-04-30

2016-07-30

2016-10-29

2016-12-29
2017-01-28
2014-07-31

2014-10-30

2015-01-29

2015-03-31
2015-05-01

2015-07-01
2015-07-31

2015-09-30

2015-12-30

2016-03-30

2016-05-30
2016-06-30

2016-08-29
2016-09-29

2016-11-29

Figure 8: Displacements measured and computed on the subvertical rod extensometer EF9, located downstream
on the bottom of the central block 17-18 at 116.3 m elevation, from July 2014 to January 2017

8
Extensómetro de fundação EF11 - Jusante
Bloco
Rogério Mota, Maria J. Coelho, J. Piteira 20-21A.
Gomes, - Cota 129,25Jorge
L. Batista, m Neves, Celso Lima, D. Silva Matos

(mm) (m)
3 240
Upwards
2 230

1 220

Water level in the reservoir


0 210

-1 200

-2 190
Downwards

-3 EF11- Rod 1 with 69.1 m (monitoring) 180

-4 EF11- Rod 2 with 19.2 m (monitoring) 170


EF11 - Rod 1 (computed)
-5 160
EF11 - Rod 2 (computed)
-6 Reservoir water level 150

-7 140
2014-07-01

2014-12-30
2015-01-29
2015-03-01
2015-03-31
2015-05-01
2015-05-31

2015-12-30
2016-01-29
2016-02-29
2016-03-30
2016-04-30
2016-05-30

2016-12-29
2017-01-28
2014-07-31
2014-08-30
2014-09-30
2014-10-30
2014-11-30

2015-07-01
2015-07-31
2015-08-30
2015-09-30
2015-10-30
2015-11-30

2016-06-30
2016-07-30
2016-08-29
2016-09-29
2016-10-29
2016-11-29
Figure 9: Displacements measured and computed on the vertical rod extensometer EF11, located downstream on
the bottom of the right bank block 20-21 at 129.25 m elevation, from July 2014 to January 2017

Good agreements were also obtained with data from other types of monitoring devices in
the dam´s body and foundation, namely displacements measured in plumb lines,
demonstrating the good performance of the dam and the adequacy of the models considered
for the analysis and interpretation of the observed behaviour during the first filling of the
reservoir and in the subsequent operation period.

5 CONCLUSION
Geological-geotechnical characterization of Baixo Sabor dam and seismic tomographic
profile by its reference line were presented. In general, seismic tomography velocity
distribution revealed a high correlation with the previous geotechnical zoning. The fair (GZ2)
to good (GZ1) quality granite corresponds to relatively higher seismic velocity zones, and
relatively lower seismic velocity zones are associated with fractured and weathered rock
(GZ3).
This seismic velocity tomography zoning allowed complementing the mechanical
characterization of the dam’s foundation. Seismic tomography also contributed to adjust the
foundation mechanical parameters for the finite element model used for the analysis and
interpretation of dam´s structural behaviour during the first filling of the reservoir and the
subsequent operation period.
The characterization of the materials’ properties (concrete and rock mass foundation) and
of the actions and responses during the first filling of the reservoir, as well as the structural
model used to simulate the structural behaviour, were presented. This model, developed by
LNEC, included a 3D representation of the dam and its rock mass foundation. It was analysed
by the finite element method, considering the actions time evolution, the foundation
deformability and the concrete viscoelastic behaviour. A good agreement between the
monitoring results and those obtained through the mathematical modelling was achieved,
which confirmed the good performance of the dam and the modelling adequacy during this
important lifetime period.
The seismic tomography method may also be used during the future dam’s exploitation
phase to evaluate the mechanical properties evolution of the rock mass foundation and so
allowing attuning the mechanical parameters for the finite element model.

9
Rogério Mota, Maria J. Coelho, J. Piteira Gomes, A. L. Batista, Jorge Neves, Celso Lima, D. Silva Matos

REFERENCES
[1] D. Demanet and D. Jongmans, Seismic tomography survey under La Gileppe dam, In
D.M. McCann et al. (eds.), Modern geophysics in engineering geology, 175-182, The
Geological Society, London, England (1997).
[2] M.J. Coelho, Crosshole seismic tomography for geotechnical investigation. (in
Portuguese), Monography, Public examination for LNEC Research Assistant, LNEC,
Lisboa, Portugal (2000).
[3] N. Plasencia, M.J. Coelho, C. Lima and L. Fialho-Rodrigues, The role of seismic
tomography in the Venda Nova II powerhouse rock mass characterization. (in
Portuguese), VII National Geotechnical Congress, Vol. I, 279-285, Porto, Portugal
(2000).
[4] M.J. Coelho, F.M. Salgado and L. Fialho-Rodrigues, The role of crosshole seismic
tomography for site characterization and grout injection evaluation on Carmo convent
foundations, In A. Viana da Fonseca and P.W. Mayne (eds.), ISC’2, 2nd International
Conference on Site Characterization, Porto, Vol. 1, 443-449, Millpress (2004).
[5] M.J. Coelho, R. Mota, A. Morgado and J. Neves, Time-lapse crosshole seismic
tomography for characterisation and treatment evaluation of the Ribeiradio dam rock
mass foundation, Submited to NUMGE 2018, June 2018, Porto, Portugal (2018).
[6] B. Lehmann, Seismic traveltime tomography for engineering and exploration
applications, EAGE, Houten, The Netherlands (2007).
[7] R. Mota and M.J. Coelho, Crosshole seismic tomographies at the foundation of Baixo
Sabor dam after the first reservoir filling. (in Portuguese), Report LNEC 209/2017,
LNEC, Lisboa, Portugal (2017).
[8] B. Butchibabu, N. Sandeep, Y.V. Sivaram, P.C. Jha, and P.K. Khan, Bridge pier
foundation evaluation using cross-hole seismic tomographic imaging, Journal of Applied
Geophysics, 144, 104-114 (2017).
[9] J.M.N.C. Pessoa, Application of tomographic techniques to crosshole seismic
exploration, (in Portuguese), Monography, Public examination for Assistant Professor,
Aveiro, Aveiro University (1990).
[10] AHBS, Geophysical study by crosshole seismic tomography – Upstream dam (in
Portuguese and in Spanish) [File AHBS_EGTSCH.EM.03.00.pdf] (2014a).
[11] AHBS, Geophysical study by crosshole seismic tomography – Upstream dam (in
Portuguese and in Spanish) [File AHBS_RelatorioTSCH_IV_06.06.14.pdf] (2014b).
[12] EDP Produção, S.A., Aproveitamento Hidroeléctrico do Baixo Sabor, Projecto. Memória
Geral e Apêndices, Volume II – Apêndices do Escalão de Montante, Tomo 1 – Estudos
Geológicos e Geotécnicos, (in Portuguese), Unpublished report, Porto (2005).
[13] N.F. Grossmann and J.P. Neves, Rock mechanics tests for the Baixo Sabor main dam,
ISRM International Symposium 2008, 5th Asian Rock Mechanics Symposium
(ARMS5), Tehran, Iran, 285-292 (2008).
[14] GEOÁREA, Geological-geotechnical report AHBS/RGG.EM.BR.01.01, Dam and
dissipation basin (foundation), Upstream dam, (in Portuguese), Unpublished report,
Baixo Sabor ACE (2015).
[15] J. Piteira Gomes, A.L. Batista and D. Silva Matos, Analysis and interpretation of the
Baixo Sabor dam’s behaviour during the first filling of the reservoir. ICOLS
Symposyum 2018, Viena, Austria (2018).

10