Documente Academic
Documente Profesional
Documente Cultură
ISRM 2003–Technology roadmap for rock mechanics, South African Institute of Mining and Metallurgy, 2003.
Abstract: The formation and mechanism of rock cavern roof arches are studied and effects of
rockbolts reinforcement are evaluated. It is concluded that a self-supported roof arch may be
formed by the excavation. To stabilize such a roof arch, fully grouted rockbolts are effective.
Tensioned cable anchors may have a negative influence on the arch stability.
Résumé: Le mécanisme et la formation de voûtes dans les cavernes rocheures sont étudiés et
l'effet du renforcement par boulonnage est évalué. Nous arrivons à la conclusion qu'une voûte
autosupportêe peut être formée par excavation. Des boulons entièrement scellés par coulis de
bêton s'avèrent efficaces pour stabiliser de telles voûtes. Les tirants d'ancrage peuvent quant à eux
nuire à la stabilité de la voûte.
549
in fact no question about an overall criterion of strength. The stability against collapse of funicular voussoir arch
Rather it is the shape of the structure that must be requires the thrust line within the 'middle third' area of the
examined. As Heyman (1995) pointed out, the stability of arch ring. For various profiles of voussoir arches, it is
the structure will be assured primarily by its shape, and not necessary to determine a minimum thickness of the arches,
at all (or only very marginally) by the strength of the which comprise the ring the thrust line does not pass
component material. Thus it is essential to design and outside. Following Inglis (1963), Irvine (1981) gave the
excavate the arch shaped roof. answer for the classical Roman voussoir arch assuming the
Under such conditions, a natural rock roof arch may be collapse mechanism illustrated in Figure 2.
formed and self-supported. Apart from the experimental
experience reported by Bergman and Bjurstron (1983), sita (degree)
t/R
Norway. The cavern has span of 61m and height of 24m. 0,08 0,08
The rock cover varies from 25m to 50m. The quality of 0,04 0,04
rock mass (Precambrian gneiss) is fair to good with Q=4 to 0 5 10 15 20 25 30 35 40
c/R
40. Horizontal to vertical stress ratio is 3.5. Numerical Figure 2 Assumed collapse Figure 3 t/R vs. c/R and t/R
studies and measurement gave the maximum roof mechanism showing hinge vs. θ relation curves
displacement of about 7mm. Due to the high horizontal in- locations and thrust line (
situ stresses some initial episodes of arching upwards are from Irvine, 1981)
found (Broch et al. 1996). 6-12m long bolts/cables with
spacing 2.5 x 2.5m were used. Broch et al. (1996) suggest a For infinite overburden (c), the general results of the
shorter bolt of 3-4m will do the support job since the minimum voussoir thickness t and the angle locating the
measured load in the bolts is very low and occurs only in a hinge at the intrados θ as a function of the radius R, are
limited length of the bolt. shown in Figure 3 and are given as
θ = sin −1 (1 / 3 ) ≅ 35.3 o (3)
Arch action t
= 2(7 − 4 3 ) ≅ 15% (4)
Arch action can be visualized by the compressive stress R
zone uniformly distributed along the arch from numerical The minimum thickness of the arch is about 15% of the
modeling. The visualization indicates the load transferring arch radius. It is noted that a threefold increase in t/R is
from the crown to the abutments and further down to the associated with only a 50% increase of θ. This indicated
sidewalls, and the compaction by which the arch is self- that the location of the hinge at the intrados varies little
supported. In classical voussoir arch theory, the action of even for widely differing forms of the arch, implying the
the arch can be described by the line of thrust. The thrust critical of the arch rib in stability. In the proceeding, the
line is a polygon representing the equilibrium of given horizontal load is not considered. From Eq.(3), with
loads. The thrust line for a general arch shape of rock increase in the horizontal in-situ stress, a thicker arch is
cavern roof is derived based on limit equilibrium analysis required. And thus a very high horizontal in-situ stress may
(Huang, 2001). Under vertical loads (q(x)) only, and not always be beneficial to the roof arch forming. A
including both the vertical and horizontal in-situ stress favorable magnitude of the horizontal stress should be
components, the thrust line of the left part of the arch is related to a certain size of the span.
Eq.(1) and Eq.(2), respectively:
Roof arch reinforcing
V W (1)
y= x − x ( x − xd ) Excavation may induce plastic/tensile zones that lead
H H to failure of the rock mass. The shear failure in weak
V Wx (2) discontinuities and shear failure due to incomplete arch
y= x− ( x − xd )
H' H' forming is common (Stille, 1992). If adequate support is not
where provided, aggressive air or water can cause further
x L
W x = ∫ γq ( x)dx ; V = W / 2 = ∫ γq ( x )dx / 2 ;
deterioration of the exposed rock. Water may build up
0 0 water pressure that increases tensile stress in
discontinuities, further dilates the joints, decreases the shear
x d = ∫ q ( x) xdx / ∫ q( x)dx ; γ = unit weight of the rock
x x
5
2,0
4 k0=0.4 1,5
k0=0.8
3
Hs /s
Hs /s
1,0
2
0,5
1
0 0,0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
L/s L/s
Phi=20 Phi=30 Phi=20 Phi=30 Phi=40 Phi=45
Figure 5 Perspective view of Figure 6 Deformed
Phi 40 Phi 45
Xiaolangdi powerhouse geometry(F1-DIANA)
Figure 4 Relations of Hs/s and L/s
strength of the rocks is in the order of 60 - 150 MPa. The
Assume a large H/s value, the variation of Hs/s with L/s lateral in-situ stress ratio (k0) is about 0.8. There are several
and φ(Phi) for k0=0.4, k0=0.8 can be plotted as shown in clay seams with thickness of 1mm to 20mm parallel with
Figure 4. As L/s increases to some extent the confining bedding planes. The existence of the clay seams became a
pressure decreases and approaches a minimum value. So major concern for the stability of the cavern crown.
that very long bolts/cables are not necessarily beneficial for Combining with other considerations such as support
551
pressure (further discussion on this issue refers to Huang et changes were taking place in the roof during
al. (2002)), this leads to a total of about 345 1500kN excavation, which means a natural roof arch may be
tensioned 25m long cable anchors installed in the machine already formed, and after pattern rockbolting the roof
hall (251.5m long, 26.2m wide, 61.4m high) roof, in arch is more stable.
addition to pattern 6-8m long rockbolts spacing 1.5m x • The cables may increase wall displacements due to the
1.5m, and 20cm wire mesh shotcrete (Wang and Solymar, mechanism of the coupling of roof deformation with
1997). the wall deformation in the cavern with high walls. The
cables cause irregular stress distribution and stress
Numerical modeling concentration in the arch. The extra upward support
force may rotate rock blocks in a way that resists the
The objective of the numerical modeling is to find a
forming of the arch, and induce sliding along joints and
rational rockbolts support design in stabilization of the
tensile stress. These turn out contrary to the
Xiaolangdi underground powerhouse roof arch and the
expectations.
walls. This is achieved by simulation of different geological
models in combination with rockbolting alternatives. Base
Table 1 Displacement (mm) calculated in the UDEC and
on the geological structures, the geological models consist
the DIANA modeling
of intact rock model (F), stratified rock mass model (E) and
jointed rock mass model (K). Rockbolting alternatives Model k0 UDEC DIANA
include no support at all (‘1’), only fully grouted rockbolts C UW DW C UW DW
(‘3’) and both tensioned cables and grouted rockbolts (‘7’). F1 0.8 6.7 19.1 22.9 5.3 16.6 19.7
For comparison and verification, both the UDEC and the F7 0.4 8.4 13.6 20.0 6.6 6.9 16.1
DIANA codes were used. F7 0.8 6.1 17.4 20.9 4.4 11.8 17.8
F7 1.0 5.5 20.1 22.1 3.4 15.4 20.8
General results Note:C=Crown,UW=UpstreemWall,DW=DownstreemWall
Response of the rock mass to the excavation:
• The maximum value of roof subsidence and
downstream wall displacement is about 5.3 - 6.7mm
and 19.7 - 22.9mm for intact rock model F1, see Table
1 and Figure 6.
• Plastic zones extend to more than 20m in the walls and
E7
to a maximum depth of 5m in the roof, see Figure 7. E3
10m-18m deep tensile zones are calculated in the Figure 7 Plastic zone in rock mass-model E3,E7 (UDEC)
walls, Figure 8(a). Only in jointed rock mass model
K1, a tensile zone of 5m deep in the roof exists.
• Figure 8(a) shows that the arch shaped roof fits the
elliptical rock ring surrounding the opening, and thus
leaves a limited range of the plastic/tensile zone. The F7F1F3 F3
high walls do not fit it. E3
• A natural roof arch with about 5m thick uniform stress
zone near the roof surface is established by the
favorable in situ stress, the good quality of rock and the
arch shape of the roof. The location and thickness of
the roof arch may be determined by use of a new
concept of invert principal stress cone.
(a) (b)
• There is the mechanism of coupling of the deformation
Figure 8 (a) Minor principal stress contour with tensile
of roof and the deformation of high walls in the cavern.
zone (red) in the rock mass with DIANA; (b) bolt load
distributions with UDEC.
Characteristics of rockbolting in stabilization of the
powerhouse cavern:
• The rockbolt should be long enough to facilitate the
• The pattern fully grouted rockbolts reduce the roof maximum shear stress that may develop and to cross
subsidence and the wall displacements, see Table1. critical joints (Figure 8(b)). The critical joint might be
Comparing with the tensioned cables, the rockbolts are defined by its critical location less than double
more effective way. Further evidences include that the thickness of the rock roof arch. When a joint is loacted
grouted rockbolts efficiently reduce the plastic zone, beyond the ctitical location, no special attension should
tensile zone and opening of the clay seams in the roof. be paid if the rock roof arch is ensured. As Figure 7
• The load built up in the rockbolts in the crown is low shows, the length of 6-8m rockbolts is long enough to
and within about 2m from the surface, see Figure 8(b), cover the plastic zone in the roof, and it is in agreement
small loads changed in the tensioned cables. The with empirical rules (Huang, 2001); while in model E7,
strength of the rockbolts is not considerably mobilized, the 25m-cables are too long to be rational.
nor the cables. These also indicate that small stress
Keystone Keystone
opening
& sliding
553
blocks is pertinent to spaces available behind the arch. It is subsidence and wall displacement. Tensioned cables may
the presence of the clay seam in the crown - a weakness be used to suspend large rock wedges or blocks locally.
zone that probably provides spaces for such an
unfavourable rotation and displacement. This reduces the Reference
thickness of the arch namely the factor of safety. Thus the
high tensioned cables take the risk to reduce rather than Bergman, Sten G.A.B and Bjurstrom, Sten.1983. Swedish
increase the factor of safety. experience of rock bolting. Proc. of the Int. Symp. On Rock
Bolting, Abisko, pp243-255.
Bjurstrom, S. 1974. Shear strength of hard rocks reinforced
Comparison between simulation and by grouted untensioned bolts. Proc. 3rd Congr. ISRM,
measurement Denver, pp.B1194-1199.
Most of the instruments in the Xiaolangdi powerhouse Broch, E, Myrvang, A. M. and Stjern, G. (1996) Support of
were installed too late during the excavation. For such a large rock caverns in Norway. Tunnelling and Underground
large and important project, the delay may simply indicate Space Technology, Vol. 11, No. 1, pp. 11-19.
that the geological conditions in the area in question were Egger, P. 1973. Rock stabilisation. Rock Mechanics,
regarded as being generally good. The first monitoring data Courses and Lectures, Udine, pp241-297
were recorded after the top heading excavation and the Gerdeen, J. C. et al. 1977. Design Criteria for Roof Bolting
installation of the cables were completed. From the data of Plans Using Fully Resin-Grouted Nontensioned Bolts ro
multiple-point borehole extensometers, rock/cable Reinforce Bedded Mine Roof, USBM, Contract No.
dynamometer, important conclusions are as follows: J=366004, Michigan Technological University, Vol. IV.
Heyman, J. 1995. The stone skeleton - structural
• During benching excavation a maximum roof
engineering of masonry architecture. Cambridge University
settlement of 2mm was recorded, much less than the
Press. 160p.
expected in the report of rock support design (YRCC,
Hibino, S., Motojima, M. and Kanagawa, T. 1983.
1994), and it confirms the results of model F.
Behavior of rocks around large caverns during excavations,
• The cavern roof has uplift movement during the first
Proc. 5th Congr. ISRM. Melbourne, pp.D199-202
benching excavation. So does the DIANA modelling of
Huang,Z. 2001. Stabilizing of rock cavern roofs by
model E7.
rockbolts. Ph.D thesis. Norwegian University of Science
• The monitored maximum displacement at lower part of and Technology, Norway.
the downstream wall, coincides with the modeling Huang,Z., Broch, E., Lu, M. 2002. Rockbolt support of
results. cavern roofs based on analytical solutions. EUROCK 2002,
• From the monitored data, it is deduced a maximum of Lisbon, Portugal
6 mm subsidence at vault and 15mm inward Inglis, Sir Charles. 1963. Applied mechanics for engineers.
displacement of the downstream wall during Dover Publications, Inc. New York.
excavation. These are close to the intact rock model F Irvine, H. Max. 1981. Cable Structures. The MIT Press.
result. Lang, T. A. and Bischoff J.A. 1984. Stability of reinforced
• Among 9, only one cable dynamometer exceeded the rock structure. In Design and performance of underground
tension load by 14% in 1998, at the same year, highest excavations (edited by Brown, E. T. and Hudson, J.A.),
load of about 79kN was recorded in bolt dynamometer pp11-18. British Geotechnical Society, London.
with others less than 20kN. These imply that small Rabcewicz, L. 1969. Stability of tunnels under rock load.
stress changes were taking place in the roof and the Water Power, 225-229 (June), 266-273(July), 297-302
cavern roof arch is self-supported. This is in (Aug.)
accordance with the numerical modeling results. Stillborg, B. 1984. Experimental investigation of steel
cables for rock reinforcement in hard rock. Ph.D. thesis.
Conclusions Luleå Univeristy, Luleå, Sweden.
Stille, H. 1992. Keynote lecture: rock support in theory and
With a certain magnitude of in-situ horizontal stress,
reasonable quality of the rock mass and arch shaped roof, a practice. Proceedings of Rock support in mining and
natural self-supported rock cavern roof arch structure may underground construction, Kaiser and McCreath (eds),
be established near the roof surface by the excavation. Balkema, Rotterdam.
Stjern, G. 1995. Practical performance of rock bolts. Ph.D.
Pattern short and closely spaced fully grouted rockbolts are
efficient to reinforce the jointed and fractured rock masses Thesis, Norwegian University of Science and Technology,
and to avoid tensile stress and sliding along the rock joints Trondheim, Norway.
Wang, X. & Solymar, Z. V. 1997. Design and Underground
in the roof, in combining with reinforced shotcrete.
In the case of the Xiaolangdi powerhouse roof where a Construction at the Xiaolangdi Mutipurpose Dam Site.
natural and reinforced roof arch has already been formed, Tunnelling and Underground Space Technology, Vol. 12,
the additional use of systematic tensioned cables may not No. 4 pp. 461-472.
Westergaard, H.M. 1964. Theory of elasticity and
be rational to the stability of the roof. The cables are neither
efficient nor safe due to inducing stress concentration and plasticity, 176p. Dover Publications, New York.
triggering the mechanism of coupling of the roof