Design aspects for concrete lined vertical shafts for hydropower

constructions

H.Wannenmacher & M. Bauert

Amberg Engineering AG. Sargans, Switzerland
F. Amann
ETH Zürich, Geological Institute, Zürich, Switzerland

Abstract: Hydropower plants are of great relevance for European energy production.
Vertical shafts in comparison to inclined shafts may serve as a cost alternative in case
of favourable ground conditions. Vertical shafts have a higher ground coverage
compared to surface parallel inclined shafts. The increased rock mass cover and
therefore higher in-situ stress magnitudes (e.g. decreasing risk for hydro-fracturing)
may allow unlined solutions instead of cost intensive steel lining (providing that water
losses are economically acceptable). Both, concrete lining and rock mass are subjected
to transient loading conditions during operation. Depending on strength and thickness
of the concrete lining, unreinforced concrete tends to crack more easily and intense
upon exceeding a critical effective internal water pressure. The structural integrity is
associated with the number of hydraulically induced cracks and the crack width at a
given internal water pressure. The design of reinforcement depends on the bedding of
the lining (e.g. deformability of the rock mass), the hydraulic conductivity of the rock
mass in relation to the concrete lining, the external and internal water pressure, and the
in-situ state of stress. The paper presents a procedure to determine the stress state
within a concrete liner and the surrounding rock mass. Furthermore, the reliance and
application of the European Standard (Eurocode EC 2) for the geotechnical and
structural design of concrete lined pressure shafts is discussed.

Theme: Design Methods

Keywords: Hydropower Construction, Shaft Lining, Eurocode, Analytical Solution

Eurock 2012 Page 1

INTRODUCTION
Modern pump storage schemes allow immediate production of peak current upon
demand. Due to the availability of modern shaft construction methods, vertical
penstocks may serve as an alternative to common inclined shaft solutions with steel
lining. In general, vertical shafts have a higher ground coverage compared to surface
parallel inclined shafts. The increased rock mass cover and therefore higher in-situ
stress magnitudes favour unlined shaft constructions instead of cost intensive steel
lining solution. Even though, economic considerations may favour the implementation
of a concrete lining in order to minimize head losses and thus allow smaller
excavation diameters and the abdication of a rock trap.
Such a decision requires a detailed ground investigation program, which is in
particular focused on in-situ stress estimates, the sensitiveness of erosion and
alteration of the rock mass due to water flow, and the hydraulic conductivity of the
rock mass to account properly for the long-term performance of the shaft during
operation. The pending decision to incorporate a concrete lining in favour of an
unlined solution often depends on the cost effectiveness of constructional aspects such
as reduced excavation diameter, minimized head losses, and maintenance of a rock
trap. Economic considerations may be in favour of a reinforced concrete lining for
pressure shaft construction. The reinforcement is hereby considered to evenly
distribute hydraulically induced cracks, and to limit crack width. The design of
concrete lining can be facilitated with either numerical approaches or analytical
solutions, which account for transient loading conditions of the concrete lining and
rock mass during operation (Schleiss 1997, Fernandez 1997). In this paper the
applicability of different analytical design approaches are compared, which allow
determining the crack development in a concrete lined pressure shaft.

VERTICAL PRESSURE SHAFT IN A DENSE ROCK MASS

A vertical pressure shaft with sufficient ground coverage was utilized to investigate
the influence of different design approaches (Table 1) on cracking of the lining. This
example is based on an actual hydropower project and was modified for the purpose of
this study. The rock mass is considered to be isotropic, homogeneous and continuous,
and thus the influence of discontinuities on the lining design is neglected. The layout
of the pump storage scheme is shown in (Figure 1); the pressure shaft is about 600 m
in length. The excavation diameter of the pressure shaft is do = 6.6 m. The lining
consists of reinforced concrete with a thickness of 30 cm. The rock mass support (e.g.
shotcrete lining) was neglected for the design consideration in this study. The
maximum transient hydraulic pressure (dynamic pressure ratio = 1.25) at the bottom
of the shaft is 75 bar.

Eurock 2012 Page 2

 SURGE SHAFT
GROUNDWATER
TABLE

PRESSURE TUNNEL

Pressure Shaft Ø = 6.0m

600m
SURGE SHAFT

CAVERN STRUCTURE

Figure 1. Longitudinal section along the water conduit

DESIGN APPROACH
In general the rock mass serves as an integral part of the lining design. The design
approach considered in this study consits of a composite 3-layer system involving the
concrete lining, the rock mass influenced by seepage flow, and the rock mass which is
not affected by seepage flow (Figure 3). All three layers of this system are faced to
various stress states which were analysed in terms of two different design approaches
(Table 1). In principle both design approaches are only relevant when the hydraulic
conductivity of the lining is lower than the hydraulic conductivity of the rock mass.

Table 1. Comparison of design approaches

Design Approach I Design Approach II

Birckenmaier (1983) EC 2 (2005)
Primary Stress State -
Secondary Stress State Feder and Arwanitakis (1976) Feder and Arwanitakis (1976)
Internal Water Pressure acting
Schleiss (1997) Schleiss (1997)
on Concrete Lining and Rock Birckenmaier (1983) EC 2 (2005)
Mass
Sum of secondary stress state and internal water pressure action
Tertiary Stress State on concrete lining and rock mass

Primary Stress State

The primary state of stress is typically influenced by the topography, the tectonic
history, landform developement and the elastic properties of the rock mass. The stress
ratio is considered to be K0 = 1 in this study. The vertical stress at the bottom of the
shaft is 21.6 MPa. The rock mass parameters considered are given in Table 2.

Eurock 2012 Page 3

Table 2. Characteristic Rock Mass Properties

Young’s modulus: Erock 34’000 [MPa]

Poisson’s ratio: νrm 0.3 [-]
Specific weight: γrm 27.0 [kN/m3]
Friction angle: ϕrm 42.0 [°]
Cohesion: crm 10.0 [MPa]
Stress ratio K0 1.0 [-]
Hydraulic conductivity: krm 8E10-7 [m/s]
Secondary Stress State
As a consequence of excavation a secondary stress state evolves. The radial and
tangential stresses were determined using the analytical solution of Feder and
Arwanitakis (1976). The total tangential and radial stresses within the rock mass are
shown in Figure 2. Although stress redistribution takes place as a consequence of the
excavation, the rock mass in this study behaves elastic due to it’s high strength.The
elastic state of the rock mass allows for a direct superposition of the subsequent stress
states during operation. In case of undesired zones of weakness the rock mass must be
treated to achieve the desired conditions with regard to deformation behaviour and
hydraulic conductivity (e.g. grouting).
45.0
Radial Stress(σ
Radial Stress (σr)
r)
40.0
Tangential Stress
Tangential Stress(σ t)
(σt)
35.0

30.0
Stress [MPa]

25.0

20.0

15.0

10.0

5.0

0.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
Distance from Tunnel Centre [m]

Figure 2.: Radial and tangential stresses after Feder and Arwanitakis (1976).

Internal Water Pressure acting on Concrete Lining and Rock Mass

The shaft lining faces a high static internal water pressure, which is superimposed
by dynamic loads during operation. When the internal water pressure exceeds the sum
of the tangential stress and the tensile strength of the concrete lining, cracking will
occureinstantaneously.
The positive effect of the external pressure (groundwater pressure) counteracting
the internal water is not considered in this study to account for the drawdown of the
watertable during construction (conservative approach). The transient recovery of the
fluid pressure in the rock mass after excavation depends on the internal water pressure,
the water losses of the pressure shaft (e.g. hydraulic conductivity of the concrete
lining) and the hydraulic properties of the rock mass. The characteristic properties of
the concrete lining including notation for the reinforcement are given in Table 3.

Eurock 2012 Page 4

Table 3. Characteristic Properties Of The Concrete Lining

Compressive strength: fck 30 [MPa]

(FS = 1.5) fcd 20 [MPa]
Tensile strength: fctk 2.9 [MPa]
(FS = 1.5) fctd 1.9 [MPa]
Young’s modulus: Ecm 33.5 [GPa]
Poisson’s ratio: νc 0.2 [-]
Nominal steel coverage: cnom 5.0 [cm]
Diameter steel bars: D 14 [mm]
Spacing: a 15 [cm]
No. reinforcement layers: n 2.0 [-]
Displacements and effective stresses within the rock mass and concrete lining
were calculated using the theory of the thick wall cylinder considering three different
layers (Schleiss, 1986). Figure 3 illustrates the three different layers with concrete
lining,  the rock mass around the tunnel which is influenced by seepage flow, and 
the rock mass which is not affected by seepage flow.

pF(R) pF(R).. Contact stress at boundary of seepage

pR
PR…….. External water pressure
pF(ro).. Contact stress acting at the joint concrete
pF(ro) to rock mass
po
u(R
)
po…… Hydraulic water pressure acting at the
u(ro
) interface of concrete to rock mass
ri

ro
R……. Radius of seepage induced zone
pi
1 R ri…..… Radius internal
2 ro……. Radius external
3
pi……. Internal water pressure
u(ro)… Displacement at the joint concrete to rock
mass
u(R)… Displacement at boundary of seepage

Figure 3.: Layered lining system (concrete lining / rock mass) under internal and external
water pressure (Schleiss, 1986)

This system comprises three unknown parameters po, pF(ro) and pF(R) (see Figure 3),
which can be determined by three equilibrium conditions as stated by Schleiss (1986).
The equilibrium conditions hereby consider:

 Equal displacement at the interface (outer radius ro) of the concrete lining and
rock mass.
 Equal displacement at the boundary at distance “R” from the tunnel, which
separates the zone affected by seepage flow from the zone not affected by
seepage flow.
 Identical water flux through the liner depending on the number of cracks
(cracks) and crack widths (2a) and the rock mass influenced by seepage flow.

qLiner 
pi  po   2a 3  cracks
12  w  w  ro  ri 
(1)

qRock Mass 
po  pR   2    k rm
w  g  ln R 

 ro  (2)

Eurock 2012 Page 5

 The determination of the number of cracks (cracks) and crack widths (2a) is based
on two different design approaches:

 Design Approach I: according to Birckenmaier (1983); this model was

suggested by Schleiss (1997) for the general design of reinforced concrete
lining.
 Design Approach II: according to EC2 (2005).

Figure 4 and Figure 5 show the effective stress in the liner and rock mass affected
by the internal water pressure considering a linear water pressure decrease in the liner
and a logarithmical water pressure reduction in the rock mass, which is influenced by
seepage flow. The effective radial stress component can be expressed by equation (3).
Crack initiation within the lining leads to an abrupt loss of the tangential stress
component accompanied by an increased hydraulic conductivity of the lining (see
table 4). The increased conductivity of the lining considers a linear water pressure
distribution along the cross section, expressed by (4).

 
2   c   po  pi    ri2  1  ro2   1  ro 
r
 r (r )   2
3  1   c  2  2  r 
 ro  ri  r  1  i 
 ro 
r 2 r2 
 pF ( ro )  2 i 2   1  o2   pF ( ro )
ro  ri  r 
(3)

pi  ro  r   po  r  ri 
p( r ) 
ro  ri
(4)

The effective radial and tangential stresses in the rock mass, which is influenced
by seepage flow were derived using equation (5) and equation (6). Assuming
axisymmetric conditions the water pressure distribution along a cross section of the
rock mass, which is influenced by seepage flow was calculated using equation (7).

p  po    ro2  1  R 2   lnR r  

 
 r (r )  R  
2  1   rm   R 2  ro2  r 2  ln R  
 
  ro  
 R2
ro2 
 pF (R )  pF ( ro )   1  2
2 
  pF (R )
R  ro  2 
r  (5)

p  po    ro2  1  R 2   lnR r   1  2  rm 

 
 t( r )  R  
2  1   rm   R 2  ro2  r 2  ln R  
  ro  
 R2 
ro2
 pF (R )  pF ( ro )    1  2   pF (R )
2 
R  ro  2
r 
(6)

 r  p  ln r r
po  ln R R



lnR 
p( r )  o

r (7)

Eurock 2012 Page 6

 12.0

10.0 Datenreihe und(σweite

Radial Stress r)

8.0 Tangential Stress (σt)

Datenreihen3
Water Pressure (p)
Datenreihen5
6.0

4.0

Stress [MPa]
2.0

0.0

-2.0

-4.0

-6.0

-8.0

-10.0

-12.0
3 3.3 3.6 3.9 4.2 4.5 4.8 5.1 5.4 5.7 6
Distance from Tunnel Centre [m]

Figure 4.: Effective stresses in the liner and rock mass, which are affected by the internal
water pressure according to Birckenmaier 1983).

12.0

10.0 Datenreihe und

Radial Stress (σrweite
)

8.0 Datenreihen3
Tangential Stress (σt)

Water Pressure (p)

Datenreihen5
6.0

4.0
Stress [MPa]

2.0

0.0

-2.0

-4.0

-6.0

-8.0

-10.0

-12.0
3 3.3 3.6 3.9 4.2 4.5 4.8 5.1 5.4 5.7 6
Distance from Tunnel Centre [m]

Figure 5.: Effective stresses in the liner and rock mass affected by the internal water
pressure (according to EC2).

The critical internal water pressure leading to systematic cracking in this study
was found to be 14.3 bar. Cracking of the lining due to shrinkage prior to watering up
as well as a temprature decrease due to watering up was not considered.
The number of hydraulically induced cracks, the crack width and average crack
spacing calculated from the two different design approaches (e.g. BM and EC2) are
summarized in Table 4. The calculation of the hydraulic conductivity of the cracked
concrete is based on the initial hydraulic conductivity of the uncracked concrete (k =
10e-8 m/s) superimposed by the hydraulic conductance of individual cracks (e.g.
number of cracks and crack width).
The calculation based on EC2 revealed a slightly higher crack spacing and a lower
crack intensity. The total tangential elongation of 5.8 mm was calculated after Schleiss
(1986) in both design approaches. The higher the crack spacing at equal elongation, as

Eurock 2012 Page 7

derived within the EC2 approach, as higher is the crack widths and the hydraulic
conductivity.

Table 4. Comparison of crack characteristics after EC2 and BM

EC2 BM Units
Number of cracks - 46 54 [-]
Average spacing between cracks sr 43.0 36.7 [cm]
Average crack width 2a 0.13 0.11 [mm]
Hydraulic conductivity -6 -6
(cracked concrete liner):
kc_cracked 4E10 3E10 [m/s]

Tertiary Stress State

The tertiary stress state evolves during watering up of the pressure shaft. Due to
the linear elastic behavior of the rock mass, a point-symmetric modeling was utilized
to calculate the tertiary stress state in this study. Figure 6 shows both, the effective
stresses in the concrete and the rock mass in vicinity of the pressure shaft. The
effective tangential stress in the rock mass is reduced by the amount of the pore
pressure.

Sec RadialStress
2nd Radial Stress…..…
(σr) Sec TangentialStress
2nd Tangential Stress(σ..t)

Ter
3rd Radial Stress--------
Radial Stress (σr) Ter
3rd tangential stress (σt)
Tangential Stress

Datenreihen5
Water Pressure (p)

45.0
40.0
35.0
30.0
25.0
Stress [MPa]

20.0
15.0
10.0
5.0
0.0
-5.0
-10.0
3 3.3 3.6 3.9 4.2 4.5 4.8 5.1
Distance from Tunnel Centre [m]

Figure 6.: Final effective stresses in the liner and the rock mass based on the combined
model of Feder & Arwanitakis (1976) with Schleiss (1997).

Eurock 2012 Page 8

INFLUENCES OF HYDRAULIC CONDUCTIVITY AND GROUND WATER TABLE
The Operational Phase
During operation, water losses through the lining highly depend on the hydraulic
conductivity of the rock mass and the concrete lining, as well as on the height of the
groundwater table. Figure 7 shows the proportional water loss within the rock mass in
comparison to the ratio of the hydraulic conductivity lining / rock mass.

100%
ri R
90% V  
Proportional Water Loss within the Rock Mass

ro  ri ro
80% V=5
V = 10
70%
V = 15
60% V = 20

50%

40%

30%

20%

10%

0%
0.001 0.01 0.1 1 10 100
Ratio of the Hydraulic Conductivity Lining / Rock Mass [-]

Figure 7.: Proportional Water Losses within Rock Mass

Within a ratio of the hydraulic conductivity lining / rock mass lower than 0.01 the
losses depend solely on the lining. Exceeding a ratio of 0.1 the losses are solely
governed by the rock mass. Within the transition zone with a ratio, ranging from 0.01
to 0.1 the dependence of the loss is a mixed mode. (Gysel 1984, Schleiss 1997,
Fernandez 1997, Seeber 1999). According to Gysel, (1984) a water loss of 1 l/skmbar
is torable for pressure shafts and tunnels. Table 5 summarizes the coherence of the
rock mass conductivity (krm= 8E10-7 m/s) considered, in regard to the groundwater table
and the conductivity of the lining. Furthermore, for each examined combination of the
lining conductivity and groundwater table, the desired rock mass conductivity for
acceptable water losses is given.

Table 5.: Compilation of water losses in addiction to concrete conductivity and groundwater
table.

Desired R.M.
Concrete Conductivity Conductivity Groundwater
Water Loss Conductivity for
Lining Concrete Rock Mass Table
accept. Water Losses
[m/s] [m/s] [m] [l/skmbar] [m/s]
cracked 3E10-6 8E10-7 0 22 3E10-8
uncracked 1E10-8 8E10-7 0 5.2 4E10-8
cracked 3E10-6 8E10-7 700 1.9 4E10-7
uncracked 1E10-8 8E10-7 700 0.4 -

Eurock 2012 Page 9

 Obviously the water losses exceed the proposed limit of Gysel (1984) in the case
of a cracked lining considering a groundwater table, and in case of a uncracked lining
considering no groundwater table.Ground improvements (e.g grouting) must be
considered to achieve the desired rock mass conductivity and thus acceptable water
losses.The desired rock mass conductivity is in the range of 3E10-8 m/s in case of a
cracked concrete lining, and 4E10-8 m/s for an uncracked concrete lining. The water
losses decrease markedly when an external groundwater table exists. Nevertheless the
criteria for the water losses are not yet met with a cracked lining, neglecting natural
sealing over time. The desired rock mass conductivity for the cracked lining is in this
case 4E10-7 m/s.

The Dewatering Phase

Any ground water pressure acting on the lining during a dewatering phase leads to
compression of the concrete lining. The compressive stress (σc) within the lining
depends on the height of the groundwater table and on the ratio of the hydraulic
conductivity of the lining / rock mass. A ratio of ~ 0.11 is demanded for the given
groundwater table of 700 m and a design strength of the concrete lining of fcd =
20 MPa in this study, to avoid damage of the concrete lining. Figure 8 illustrates the
design process, which results in a minimum hydraulic conductivity of 1E10-7 m/s,
which is only met by a cracked concrete lining. Controlled cracking of the concrete
lining is therefore demanded to avoid substantial failure of the shaft lining during a
dewatering phase.

ri R
  10
ro  ri ro

Figure 8.: Design Chart for Dewatering of Concrete Lined Shaft

Eurock 2012 Page 10

DISCUSSION
As soon as the internal water pressure exceeds the sum of the tangential stresses
and the tensile strength in the concrete lining, cracks will be initiated. The hydraulic
conductivity of the lining increases as a consequence of systematic cracking, but is of
subordinate relevance, since the hydraulic conductivity of the rock mass and the actual
groundwater table fluctuation govern the overall water losses from the shaft.
Nevertheless, the methods for computing crack development are of importance for the
problem statement, especially in case of an overall drawdown of the groundwater
table, a very low groundwater table and in case of a very low hydraulic conductivity of
the rock mass. The reinforcement is relevant for the lining stability to in terms of a
limited crack development (e.g. number of cracks and crack width). The design
process of a concrete lined shaft demands for cracked lining to avoid substantial
failure of the shaft lining during a dewatering phase. Nevertheless, the crack
development must be balanced to comply with acceptable water losses during
operation.

REFERENCES

Birckenmaier, M. 1983. Über Nachweise im Gebrauchszustand, Schweizer Ingenieur und

Architekt , No6.
Feder, G. and Arwanitakis, M. 1976. Zur Gebirgsmechanik ausbruchsnaher Bereiche
tiefliegender Hohlraumbauten (unter zentralsymmetrischer Belastung), Berg- und
Hüttenmännnische Monatshefte, H.4
Fernandez G. 1994. Behavior of Pressure Tunnels and Guidelines for Liner Design, Journal of
Geotechnical Engineering. Vol 120, No.10
Gysel, M. 1984. Bestimmung der Felsdurchlässigkeit aufgrund von Stollenabpressversuchen,
Wasser, Energie, Luft 76, H. 7/8.
Schleiss, A. 1997. Design of reinforced concrete linings of pressure tunnels and shafts,
Hydropower and Dams, Issue Three.
Schleiss, A. 1985. Bemessung von Druckstollen, Teil I, Literatur, Grundlagen, Felshydraulik
insbesondere Sickerströmungen durch Auskleidung und Fels. ETHZ, VAW, Mitteilung 78
Schleiss, A. 1986. Bemessung von Druckstollen, Teil II, Einfluss der Sickerströmung in
Betonauskleidung und Fels, mechanisch-hydraulische Wechselwirkung,
Bemessungskriterien. ETHZ, VAW, Mitteilung 86
Seeber G. 1999: Druckstollen und Druckschächte – Bemessung– Konstruktion – Ausführung.
Stuttgart: Enke im Thieme Verlag.

Eurock 2012 Page 11