Geotechnical and Geological Engineering 19: 21^42, 2001.

21
# 2001 Kluwer Academic Publishers. Printed in the Netherlands.

Kinematic and Block Theory Analyses for Shiplock

Slopes of the Three Gorges Dam Site in China

JEONGI-GI UM AND PINNADUWA H. S. W. KULATILAKE

Department of Mining and Geological Engineering, University of Arizona, Tucson, AZ 85721,
USA
(Received 1 February 2000; revised 4 January 2001; accepted 9 March 2001)

Abstract. The granitic rock mass that exists in the shiplock region of the Three Gorges dam site
contains a number of major discontinuities and about four sets of minor discontinuities. One
hundred and thirty three major discontinuities have been mapped around the shiplock covering
an area of 1740600 m. These major discontinuities were used to perform rock slope kinematic
and block theory analyses. Kinematic analyses were performed under the following two cases:
(1) assuming all the mapped discontinuities cross the shiplock; (2) using only discontinuities that
actually intersect the shiplock. Under case (1) and case (2) the shiplock faces in the proposed
permanent shiplock region in fresh rock were found to be stable up to a cut slope of about
45 and 58, respectively. Block theory was applied to identify different block types that exist
on the shiplock faces and to estimate the maximum safe slope angles on the shiplock faces.
The orientations of the major discontinuities that actually intersect the shiplocks were considered
in this analysis.The total length of the shiplock (1750 m) was divided into 50 m segments. From the
stereo-plots, the key blocks (Type I) and/or potential key blocks (Type II) were found for only ¢ve
segments of the shiplock slopes. It was found that the dip of the cut slope should be less than about
60 to avoid creation of a key block on the proposed shiplock slopes. However, it is important to
keep in mind that these conclusions are based on the kinematic analyses performed using only
the major discontinuities. Further kinematic as well as kinetic analyses are recommended incor-
porating minor discontinuities, water forces, earthquake forces etc. before making the ¢nal con-
clusions about the maximum safe slope angle for the shiplock region.
Key words: block theory, kinematic analysis, rock slope stability, shiplocks, Three Gorges pro-
ject.

Introduction
The Yangtze River is the largest river in China. From its headwaters to its estuary,
the Yangtze River meanders over 6300 km (Figure 1) and its annual runoff into
the sea amounts nearly 1000 billion m3 (Ha, 1993). Its total drop is more than 5800 m
with a water power potential of up to 268,000 Mw. In order to harness the river and
develop its water resources, extensive efforts in investigation, planning, design and
scienti¢c research have been suggested in order to exploit the great water power
potential of the Yangtze River (Chen, 1986).
The river stretches from Fengjie to Yichang, about 200 km long, and cuts through
three majestic canyons Qutang Gorge, Wuxia Gorge and Xiling Gorge known as the
22 J-G. UM AND P. H. S. W. KULATILAKE

Figure 1. Geographical positions of Yangtze River and Three Gorges project in China.

Three Gorges. The Three Gorges Project is located in the middle of Xiling Gorge in
Yichang, Hubei Province. The project has tremendous multi-purpose bene¢ts such
as £ood control, electric power generation, navigation, irrigation, tourism and
¢shery. The project is composed of concrete dams, £ood discharging facilities, power
houses, sluice outlets, and navigation structures (Figure 2). The project construction
began in October 1993. The total estimated construction time for the project is
18 years. For further details related to the project construction, the reader is referred
to the report by Ha (1993).
The downstream axis of the permanent shiplock makes 111 degrees with the north
direction. The depth of rock excavation needed to construct the permanent shiplocks
is high with a maximum of 170 m in some locations. Operational conditions of the
shiplocks demand steep rock slopes on both sides of the shiplocks which are about
1600 m long. The granitic rock mass which exist in the shiplock region contains
a number of major discontinuities (length greater than about 20 m), and a number
of sets of minor discontinuities (length less than about 20 m) (Figure 3). Therefore,
the rock engineering problems of high steep rock slopes in the shiplock region
are very complicated. Rock mass failures in the shiplock region must be fully
prevented to assure safe travel of people in the ships and £ood-free conditions
downstream. Therefore, it is important to devise a design scheme to provide stable,
steep, high rock slopes in the shiplock region. To achieve this ultimate goal, as
 ANALYSES FOR SHIPLOCK SLOPES
23

Figure 2. Layout of the Three Gorges Project (after Ha, 1993).

 24
J-G. UM AND P. H. S. W. KULATILAKE

Figure 3. Major discontinuities in the proposed permanent shiplock area.

ANALYSES FOR SHIPLOCK SLOPES 25

the ¢rst step, discontinuity network in the rock mass close to the proposed perma-
nent shiplock region was characterized (Kulatilake et al., 1996; Kulatilake, 1998).
This was done using about 2050 discontinuity trace data as mapped on the walls
and the roof of a 400 m tunnel which is located close to the proposed shiplock region.
This paper describes the second step performed towards achieving a design scheme,
but further research is still necessary.

Geology
Comprehensive geological research for the Three Gorges Region has been in pro-
gress since the 1950s and is still continuing. The dam site of the Three Gorges Project
is situated in the south part of the Huanglin anticline. The Huanglin anticline is
72 km long along the north^south (NS) direction and 35 km wide. It consists of
crystalline rock masses of the Presinian system. The major rock type is Huanglin
granite (hornblend-biotite-plagioclase granite) which is about 90% of the total area
of the dam site, with the rest being diorite. There are many intrusive dikes in
the rock mass. The fresh granite and diorite are hard with saturated compressive
strengths around 100 MPa and 140 MPa, respectively. The ground surface has been
seriously weathered. The thicknesses of weathered layers are different in ridges and
ravines with an average thickness of 20^40 m. According to the degree of
weathering, the rock mass is classi¢ed into 4 types (Wang, 1986): (a) completely
weathered; (b) strongly weathered; (c) weakly weathered; and (d) slightly weathered
and fresh rock. In the shiplock area itself, the major rock type is again granite with
the rest being diorite. Diabase, quartzose and pegmatite can be found in the dikes
or reefs. Also, Horneblende Schist lenses can be found.

Major Discontinuities (Faults and Dikes) in the Proposed Shiplock

Region
One hundred and thirty three major discontinuities have been mapped around the
permanent shiplock covering an area of 1740600 m (Teng 1994). Figure 3 shows
the locations as well as the strike directions for these major discontinuities.
Figure 4 shows the pole distribution of these major discontinuities on a lower
hemispherical equal-area polar diagram. High orientation variability of the major
discontinuities is well depicted by the polar diagram. Predominant strike directions
of the major discontinuities seem to be along ENE^WSW and NW^SE directions.
Most of them have steep dip angles varying between 60 and 70 . The persistence
of most of these major discontinuities is expected to be between 100 m and 300 m.
The friction angle values fj , of the major discontinuities obtained through in-situ
direct shear tests range from 30 to 40 (Xuecheng 1993; Teng 1994). To obtain
conservative results, the minimum value of fj of 30 was used in the performance
of kinematic and block theory analyses.
26 J-G. UM AND P. H. S. W. KULATILAKE

Figure 4. Orientation distribution of major discontinuity poles in the shiplock area on a lower hemisphere
equal area polar diagram.

Kinematic Analyses
``Kinematic'' refers to the motion of bodies without reference to the forces that cause
them to move (Goodman, 1989). Kinematic analyses are very useful to investigate
possible failure modes of rock masses which contain discontinuities. Failure
involving movement of rock blocks on discontinuities combine one or more of
the three basic modes: plane sliding, wedge sliding and toppling. For the proposed
permanent shiplock region, kinematic analyses were performed to estimate maxi-
mum safe slope angles with respect to the aforementioned three basic failure modes.
The basic concepts related to estimation of maximum safe slope angles for the three
basic modes of failure are discussed by Goodman (1989).

Block Theory Analysis

The principal idea behind block theory analysis (Goodman and Shi 1985) is that it
allows many different combinations of discontinuities to be passed over and to
directly identify and consider critical rock blocks known as ``key blocks''.
Figure 5 shows ¢ve types of blocks in a surface excavation formed by discontinuities.
Although the actual blocks are in three dimensions, to simplify the illustration a
two-dimensional ¢gure is used. Types of blocks can be divided into in¢nite and ¢nite
blocks. An in¢nite block (type V), as shown in Figure 5(a) is not dangerous as long as
it is incapable of internal cracking. Finite blocks can be categorized into
ANALYSES FOR SHIPLOCK SLOPES 27

Figure 5. Blocks in a surface cut. (a) in¢nite; (b) tapered; (c) stable; (d) potential key block; (e) key block.

non-removable and removable blocks. Figure 5(b) is an example of type IV

non-removable tapered blocks. It is ¢nite, but it cannot come out to free space
because of its tapered shape. Finite and removable blocks can be separated into
three categories, namely type III, type II, and type I, and identi¢cation of these
blocks plays an important role in rock slope design. As shown in Figure 5(c), a
type III block is stable without friction under its gravity alone. A type II block which
is in Figure 5(d) can be remained stable as long as the sliding force on the block is less
than its frictional resistance. Under only gravitational loading, type II blocks are
stable. However, it can come out into the free surface of excavation if there are
external forces like water forces, inertia forces etc. that make the total sliding force
greater than the frictional resistance. Therefore, type II blocks are also called poten-
tial key blocks. Finally, a key block which is denoted by type I shown in Figure 5(e)
can slide into free space under gravitational loading without any external force
unless a proper support system is provided. Therefore, the identi¢cation of key
blocks is one of the most important parts in a rock slope stability analysis.

STEREOGRAPHIC PROJECTION OF A DISCONTINUITY PLANE

The stereographic projection is a strong tool for analysis of three dimensional struc-
tures such as rock block geometry. Figure 6 shows the stereographic construction of
a discontinuity plane on upper hemisphere. Ui and Li denote the upper and lower half
spaces of the discontinuity plane Pi , respectively, and its half spaces can be rep-
resented by binary digits. The number 0 corresponds to the symbol Ui which is
the half space above Pi and the number 1 corresponds to the symbol Li which is
the half space below Pi . In the upper hemisphere projection, the region above a
discontinuity plane …Ui † is the area within the great circle of discontinuity plane
Pi , and the region below the plane …Li † is the area outside of its great circle as shown
in Figure 6.
28 J-G. UM AND P. H. S. W. KULATILAKE

Figure 6. Stereographic construction of a discontinuity plane on upper hemisphere.

THE FINITENESS AND REMOVABILITY OF BLOCKS

Figure 7 provides some terminology related to the representation of an excavation

surface, on a stereonet. The rock mass side from an excavation surface is termed
the Excavation Pyramid (EP) on the stereonet and it occupies the outside region
of the great circle corresponding to the excavation surface. The free space region
from an excavation is termed the Space Pyramid (SP) on the stereonet and it falls
inside the great circle corresponding to the excavation surface. The spherical regions
formed on the stereonet through the intersection of discontinuities are known as
Joint Pyramids (JP). According to the ¢niteness theorem (Goodman and Shi, 1985),
a block is ¢nite if it satis¢es the criterion given below (that is JP is fully within SP):

Joint Pyramid (JP)  Space Pyramid (SP) …1†

Figure 7. Stereographic construction of an excavation plane on upper hemisphere.

ANALYSES FOR SHIPLOCK SLOPES 29

The blocks which do not satisfy the criterion given above falls into the category of
in¢nite blocks (type V). Therefore, by plotting discontinuity planes and the exca-
vation surface on a stereonet and using the criterion given above, the ¢nite blocks
can be separated from the in¢nite blocks.
According to the removability theorem (Goodman and Shi 1985), for a block to be
removable, in addition to satisfying the criterion given in Equation (1), the JP cor-
responding to the block should be non-empty. If JP is empty and the block satis¢es
the criterion given in Equation (1), the block belongs to the ¢nite non-removable
(tapered) category (type IV). Since JP is empty for ¢nite tapered blocks, they do
not show up on a stereographic projection. Therefore, the joint pyramids that satisfy
the criterion given in Equation (1) on a stereographic projection belong to ¢nite
removable block categories (types I, II and III).

SEPARATION BETWEEN TYPES I, II AND III BLOCKS

Failure involving movement of rock blocks on discontinuities combine one or more

of the following three basic modes: (a) lifting; (b) sliding on a single plane; and
(c) sliding on two planes. Only one joint pyramid exist for each of the above
mentioned movement directions. A mode analysis (Goodman and Shi 1985,
Chap. 9) can be performed to ¢nd JPs corresponding to all the movement directions.
Type III blocks lack a movement mode. On the other hand, a movement mode can be
found for each of the type I and II blocks. This concept allows separation of type III
blocks from types I and II blocks.
Equilibrium equations given in Chapter 9 of Goodman and Shi (1985) can be used
to compute sliding forces, F, under the resultant active force, r, for the modes of
lifting, single plane sliding and double plane sliding. A positive F corresponds to
a type I block. A type II block produces a negative F. The authors have developed
a computer code called SFORCE to perform this sliding force analysis.

Results of Kinematic Analyses

Kinematic analyses were conducted using the mapped major discontinuities in the
shiplock area to estimate maximum safe slope angles with respect to possible plane
sliding, wedge sliding and toppling failure. Each major discontinuity was treated
as a single feature. A computer code KINEM was developed to calculate maximum
safe slope angles. Although the mapped major discontinuities are shown in
Figure 3 (Teng 1994), some uncertainty exists concerning the actual extent of these
discontinuities. Therefore, the kinematic analyses were conducted under two differ-
ent cases:
(1) assuming all mapped discontinuities cross the shiplock;
(2) using only discontinuities that actually intersect the shiplock.
30 J-G. UM AND P. H. S. W. KULATILAKE

For case 1, orientation of all the major discontinuities which appear on Figure 3 were
considered in the kinematic analyses. Under this case, indirectly, it was assumed that
any major discontinuity appearing on Figure 3 has a possibility to cross the proposed
permanent shiplock. Thus, results from these analyses can be considered to be on the
conservative side. The major discontinuities that at least intersect one of the pro-
posed permanent shiplocks according to Figure 3 were used in the kinematic analyses
in the second case. Out of all the discontinuities that can be seen on Figure 3, about
25% were found to intersect at least one of the shiplocks. If the major discontinuity
map given in Figure 3 is reliable, then these analyses should provide results closer
to the reality.

FOR CASE 1

Plane Sliding. The maximum safe slope angle obtained for each major discontinuity
is shown in Figure 8(a) and (b), respectively for cut strike directions 111 
(South^West shiplock face) and 291 (North^East shiplock face). The location
of each bar shown in Figure 8 corresponds to the actual location of the major dis-
continuity in the shiplock area. The starting and ending points of each bar represent
starting and ending X coordinates of each major discontinuity. Note that the X axis
coincides with the strike of the shiplock faces. Therefore, the location of each major
discontinuity and the maximum safe slope angle corresponding to it can be identi¢ed
from this diagram. Most of the bars in Figure 8(a) and (b) indicate maximum safe
slope angles between 65 and 90 . The results obtained for the whole shiplock region
(disregarding the actual locations of major discontinuities) are shown through
histograms in Figure 9(a). Most of the maximum safe slope angles were found
to be greater than 65 .

Wedge Sliding. For the wedge sliding mode, only the discontinuities which are
located within a particular distance were considered in calculating the maximum
safe slope angles. Results for individual wedges are shown in Figure 8(c) and (d).
Similar to Figure 8(a) and (b), each bar represents starting and ending X coordinates
of each wedge. Figure 9(b) shows the histograms obtained for the maximum safe
slope angle for the whole shiplock region disregarding the actual locations of
the wedges in the shiplock area. For most of the wedges, maximum safe cut slope
angles greater than 45 were obtained.

Toppling. Toppling failures usually occur in layered rock such as slate, schist, and
sedimentary rocks. The chance to form thin layers in a granitic rock mass is
extremely low except in sheeted granite. The granitic rock mass that exist in the
shiplock region does not belong to sheeted granite. Therefore, it is not really necess-
ary to consider the toppling mode for the shiplock region. Anyway, to be on the
conservative side, in this study, kinematic analysis was conducted for toppling.
Results for individual discontinuities are shown in Figure 8(e) and (f).
ANALYSES FOR SHIPLOCK SLOPES 31

Figure 8
32 J-G. UM AND P. H. S. W. KULATILAKE

Figure 8. Maximum safe slope angles along shiplocks resulting from kinematic analyses conducted for
Case 1.
ANALYSES FOR SHIPLOCK SLOPES 33

Figure 9. Histograms of maximum safe slope angles for di¡erent failure modes, based on kinematic analy-
ses conducted for Case 1.
34 J-G. UM AND P. H. S. W. KULATILAKE

Figure 9(c) shows the histograms obtained for the maximum safe slope angle for the
whole shiplock region disregarding the actual locations of the major discontinuities
in the shiplock area. Most of the maximum safe slope angles for the toppling case
were found to be greater than 85 .

FOR CASE 2

The total length of the shiplock (1759 m) was divided into 50 m segments. For each
of these segments, ¢rst block theory analysis (Goodman and Shi, 1985) was
performed, using only the major discontinuities that at least intersect one of the
proposed permanent ship locks, before peforming the kinematic analysis. Those
segments which did not contain either a key block or a potential key block were
not subjected to kinematic analyses. The kinematic analyses, therefore, were
performed only for ¢ve segments where either key blocks or potential key blocks
were found. Table 1 gives the major discontinuities and their orientations for
the segments which were subjected to kinematic analyses. Note that analyses were

Table 1. Major discontinuity orientation data for some selected segments of the shiplock
Dip
Location (m) Discontinuity Dip Direction Comments

15150^15200 1. bu1020 82 316 Strike of SW Shiplock Face: 111

2. F1062 60 326
3. F1061 75 351
4. F1067 80 046

15550^15600 1. F203 70 264 Strike of NE Shiplock Face: 291

2. F204 82 045
3. bu1080 80 331
4. F1035 84 285
5. F8 80 247

15750^15800 1. bu1001 78 300 Strike of SW Shiplock Face: 111

2 F1005 80 221
3. bu1002 80 325
4. bu1003 50 320

16080^16130 1. F205 65 269 Strike of SW Shiplock Face: 111

2. F1009 80 274
3. F1011 70 101
4. bu1007 85 329

16420^16470 1. F1021 80 306 Strike of NE Shiplock Face: 291

2. F1023 76 078
3. F1024 70 251
4. bu1009 78 090
F: fault, bu: dyke.
Orientation of Cut Slopes: Strike of Cut (SW shiplock face) ˆ 111 ; Strike of Cut (NE shiplock
face) ˆ 291 ; Dip Direction of SW shiplock faceˆ 201 ; Dip Direction of NE shiplock face ˆ 21 .
ANALYSES FOR SHIPLOCK SLOPES 35

performed separately for the SW shiplock face (strike of cut 111 ) and for the NE
shiplock face (strike of the cut 291  ). The maximum safe cut slope angles obtained
for plane sliding, wedge sliding and toppling modes for these different segments
are given in Table 2.

Results of Block Theory Analysis

As mentioned before block theory analysis was performed for each 50 m segment
using the major discontinuities shown in Figure 3 that at least intersect one of
the proposed permanent ship locks. Maximum safe slope angles were estimated with
respect to possible single plane sliding, double plane sliding and free falling. The
maximum safe slope angle corresponding to a possible mode of failure can be con-
sidered as the maximum dip of a cut slope which will not allow formation of a type
I block under the mode considered. Analysis was done separately for the SW
shiplock face (strike of cut 111 ) and for the NE shiplock face (strike of cut 291 ).
First, removable blocks (types I, II and III) were identi¢ed for each segment. Then
mode analysis was performed to separate type III blocks from types I and II blocks.
Out of all the segments analyzed, only ¢ve segments produced type I or type II
blocks. Finally, type I blocks were separated from type II blocks for each segment
analyzed.
Figure 10 shows the identi¢cation of removable blocks for the section
16420^16470 (m) in the shiplock region using the stereographic projection. The
orientations of the discontinuities and the excavation surface for the sections are
listed in Table 1. When the dip of the excavation surface is between 0 and 58 ,
the JPs 0100, 0000, and 0001 provide removable blocks. The JPs 0010, 0000,
and 0001 provide removable blocks when the dip of the excavation surface is between
58 and 62 . For the dip of the excavation between 62 and 75 , the JPs 0011, 0010
and 0001 correspond to removable blocks. Finally, for excavation dips between 75 ,
and 90 , the JPs 0011, 0010 and 1011 provide removable blocks.
The mode analysis conducted for the section 16420^16479 (m) with strike of cut
291 (dip of the cut is 90 from the strike in the clockwise direction) showed that
JPs 0100, 0000 and 0001 correspond to type III blocks. Table 3 shows the results
obtained for F using the discontinuity data of same section when active force r
is only due to the weight of block (due to gravity alone). The results in Table 3
allowed separation between type I and type II as shown in Table 4. Table 4 shows
the maximum safe slope angle (MSSA) for all identi¢ed types I and II blocks. This
table also shows the maximum safe slope angle obtained for the segments that pro-
duced either a type I or type II block. It is important to note that when the cut
slope angle is less than MSSA, both types I and II blocks become in¢nite (type
V) blocks. To calculate the required support, limit equilibrium analysis can be used
incorporating the selected slope value, external forces acting on the block and
the weight of the block.
36 J-G. UM AND P. H. S. W. KULATILAKE

Figure 10. Identi¢cation of removable blocks for section 16420^16470 (m) of the shiplock using
stereographic projection.

Note that the block theory analysis results for each of the ¢ve segments are based
on analyzing either four or ¢ve discontinuities together (Table 4). For segments with
four discontinuities, the block theory analysis can be performed four times by taking
three discontinuities at a time (discontinuity combinations 1^2^3, 1^2^4, 1^3^4 and
2^3^4). However, the results with respect to the sliding modes corresponding to
type I blocks, sliding modes corresponding to type II blocks and MSSA would
be exactly the same as that for the analysis using all four discontinuities together
at one time. The only difference would be the individual blocks belonging to
type I or type II. Each analysis that uses three discontinuity planes would produce
tetrahedral blocks when the three discontinuities are combined with the excavation
surface. Polyhedral blocks with ¢ve faces are formed when four discontinuities
are combined with the excavation surface. For segments with ¢ve discontinuities,
the block theory analysis can be performed ¢ve times by taking four discontinuities
at a time as well as 10 times by taking three discontinuities at a time. The results
with respect to the sliding modes corresponding to type I blocks, sliding modes cor-
responding to type II blocks and MSSA would be the same as that for the analysis
using all ¢ve discontinuities together at one time. Again the only difference would
be the individual blocks belonging to type I or type II. Polyhedral blocks with
Table 2. Maximum safe slope angles for di¡erent failure modes in di¡erent segments of the shiplock region resulted from Case 2 kinematic analyses
Location
(m) D1 D2 D3 D4 D5 T1 T2 T3 T4 T5 I12 I13 I14 I15 I23 I24 I25 I34 I35 I45

15150^ 90 90 90 90 ööö 90 90 90 46 ö 90 90 90 ö 80.7 90 ö 90 ö ö

15200
15550- 90 82.7 83.5 90 90 90 90 90 90 90 65.8 90 90 67.4 79.4 78 65.2 82.5 90 90
15600
ANALYSES FOR SHIPLOCK SLOPES

15750- 90 80.6 90 90 ö 90 90 90 90 ö 84.1 89.7 90 ö 84.9 90 ö 90 ö ö

15800
16080- 80.1 87 90 90 ö 90 90 90 90 ö 90 90 71.7 ö 90 86.1 ö 90 ö ö
16130
16420- 87.4 82.3 90 85.6 ö 90 90 90 90 ö 62.4 90 57.8 ö 90 75.3 ö 90 ö ö
16470
D1 : For Plane Sliding Along Discontinuity 1 I12 : For Sliding on Planes 1 and 2 Along I12
D2 : For Plane Sliding Along Discontinuity 2 I13 : For Sliding on Planes 1 and 3 Along I13
D3 : For Plane Sliding Along Discontinuity 3 I14 : For Sliding on Planes 1 and 4 Along I14
D4 : For Plane Sliding Along Discontinuity 4 I15 : For Sliding on Planes 1 and 5 Along I15
D5 : For Plane Sliding Along Discontinuity 5 I23 : For Sliding on Planes 2 and 3 Along I23
T1 : For Toppling on Discontinuity 1 …fj ˆ 35 † I24 : For Sliding on Planes 2 and 4 Along I24
T2 : For Toppling on Discontinuity 2 …fj ˆ 35 † I25 : For Sliding on Planes 2 and 5 Along I25
T3 : For Toppling on Discontinuity 3 …fj ˆ 35 † I34 : For Sliding on Planes 3 and 4 Along I34
T4 : For Toppling on Discontinuity 4 …fj ˆ 35 † I35 : For Sliding on Planes 3 and 5 Along I35
T5 : For Toppling on Discontinuity 5 …fj ˆ 35 † I45 : For Sliding on Planes 4 and 5 Along I45
37
38 J-G. UM AND P. H. S. W. KULATILAKE

Table 3. Net sliding force (F) for the section 16420^16470 (m) with strike of cut
291 under gravity alone
Sliding Planes Net Sliding Force (F)

None 1.00 W
1 0.86 W
2 0.80 W
3 0.70 W
4 0.83 W
1, 2 0.16 W
1, 3 0.65 W
1, 4 ÿ0.19 W
2, 3 ÿ2.10 W
2, 4 ÿ0.16 W
3, 4 ÿ1.41 W

six faces are formed when ¢ve discontinuities are combined with the excavation
surface.
It is important to note that under only gravitational loading MSSA corresponding
to each type II block is 90 . However, by assuming that the resultant force of weight
of the block and possible external forces acting on the block exceeds the frictional
resistance, type II blocks gave rise to many very low maximum safe slope angles.
Many in fact below the basic friction angle, fb , of 30 and one even at 10 . Note
that MSSA values below the basic friction angle are only possible with type II blocks
in the presence of external forces, and never with type I blocks. For types I or II
blocks, to have a design slope angle greater than MSSA, it is necessary to perform
a limit equilibrium analysis incorporating all the forces to estimate the required
external rock support system.

Comparison Between the Results of Kinematic and Block Theory Analyses

Maximum safe slope angles for type I blocks correspond very well to the same sliding
modes found in Case 2 of the kinematic analysis (compare Tables 2 and 4). Both
methods show that the lowest maximum safe slope angle for type I blocks for
the region investigated is 62 . For the type II blocks under gravitational loading,
irrespective of the plunge angle of the dip vector (for single plane sliding) or the
line of intersection (for the wedge mode) the sliding force due to gravity is less than
the frictional resistance. Therefore, the MSSA can be 90 . According to the proce-
dure used in the kinematic analysis, the dip vectors and lines of intersections having
a plunge angle less than the friction angle produce maximum safe slope angles
of 90 . In the case of plane sliding under gravitational loading, the dip vectors having
a plunge angle greater than the friction angle of the discontinuity do not produce
type II blocks. However, in the case of wedge sliding under gravitational loading,
it is possible to have type II blocks having the plunge of line of intersection exceeding
both friction angles of the discontinuities forming the wedge. These blocks under
Table 4. Results of block theory analysis for di¡erent angles of the cut slopes for some selected segments of the shiplock region
Segment (m) Dip of cut Type I Blocks Type II Blocks
Strike of Cut slope …a † Key blocks Potential key blocks Type III Blocks MSSA

15150^15200/ 0^80 1001, 1010, 1011 90

111 80^90 0011…S23 †

15550^15600/ 0^65 10110, 00110, 00111, 78.1

291 00010, 00011, 00100
65^66 00000…S25 † 00110, 00111, 00010,
00011, 00100
66^67 00000…S25 †, 00001…S12 † 00110, 00111, 00010,
00011, 00100
67^78.1 00001…S12 †, 01001…S15† 00011
ANALYSES FOR SHIPLOCK SLOPES

78.1^79.5 10001…S24 † 00001…S12 †, 01001…S15 † 00011

79.5^82.5 10001…S24 †, 11011…S23 † 00001…S12 †, 01001…S15 † 00011
82.5^90 10001…S24 †, 11011…S23 †, 10011…S34 † 00001…S12 †, 01001…S15 † 00011

15750^15800/ 0^10 1000, 1110, 1010 81

111 10^81 0010…S34 † 1000, 1110, 1010

81^85 1011…S2 † 0010…S34 † 1000, 1110, 1010

85^89.7 1011…S2 †, 0011…S12 † 0010…S34 † 1010
89.7^90 1011…S2 †, 0011…S12 †, 0111…S13 † 0010…S34 † 1010

16080^16130/ 0^14 1000, 0000, 0100 72

111 14^15 0001…S23 † 0000, 0100
15^16.5 0001…S23 †, 0101…S13 † 0100
16.5^72 0001…S23 †, 0101…S13 †, 0111…S12 †

72^90 0011…S14 † 0111…S12 †

16420^16470/ 0^58 0100, 0000, 0001 62

291 58^62 0010…S14 † 0000, 0001

62^75 0011…S12 † 0010…S14 † 0001

75^90 0011…S12 † 0010…S14 †, 1011…S24 †

MSSA: Maximum Safe Slope Angle (degrees).

Si : Sliding on a single joint face (i) Sij : Sliding on two faces (i and j).
39
40 J-G. UM AND P. H. S. W. KULATILAKE

kinematic analysis produce maximum safe slope angles less than 90 . For these
blocks it is not possible to make a comparison between the results obtained through
kinematic and block theory analysis. For the rest of the type II blocks, the results
agree between the two analyses. It is important to note that kinematic analysis pro-
duce maximum safe slope angles less than 90  for types IV and V blocks. However,
with respect to block theory analysis the maximum safe slope angle corresponding
to types IV and V blocks can be considered as 90 because they do not provide
removable blocks. Block theory does not account for toppling failure, thus no com-
parison is possible.
The primary advantage of block theory over traditional kinematic analyses is that
it gives the ability to identify the keyblocks that require immediate attention. It
separates the most important and dangerous blocks from the less critical ones.
However, kinematic analysis should be performed in conjunction with block theory
to look into toppling failure.
Both methods contain assumptions that lead to both conservative and
non-conservative results. The assumption that all discontinuities are of in¢nite
extent leads to conservative results. Leaving many discontinuities out and not con-
sidering repeating joint sets, on the other hand, leads to non-conservative results.
By far the criterion of most consequence is that this study assumed only gravitational
loading as the force which contributes to sliding. External forces, such as water
forces and earthquake forces need to be considered in a future study.

Conclusions
Kinematic analysis for plane sliding has resulted in maximum safe cut slope angles
greater than 65 for most of the discontinuities. Kinematic analysis has produced
maximum safe cut slope angles greater than 45  for most of the wedges under
Case 1 (where it is assumed that all mapped discontinuities cross the shiplock).
Under Case 2 (using only discontinuities that actually intersect the shiplock), for
wedges, the maximum safe cut slope angles were found to be greater than 57  .
For the wedge sliding mode, only the discontinuities which are located within a
50 m distance were considered in obtaining the maximum safe slope angles. Maxi-
mum safe cut slope angles greater than 85 were obtained for most of the dis-
continuities in the toppling case. According to the results obtained for both
Cases 1 and 2, it seems that the shiplock faces in the proposed permanent shiplock
region in fresh rock are stable up to a cut slope of about 45  .
For the block theory analyses performed, only the discontinuities which are
located within a 50 m distance were considered in obtaining the maximum safe slope
angles. Note that the block theory analysis was conducted only for the dis-
continuities under Case 2. The key blocks (type I) and/or potential key blocks
(type II) were found only for ¢ve segments of the shiplock slopes. The dip of
the cut slope should be less than about 60 to avoid creation of a key block on
the proposed shiplock slopes.
ANALYSES FOR SHIPLOCK SLOPES 41

It is interesting to note that the shiplock slopes constructed so far have about 60
slopes at the mid height level in slightly weathered or fresh rock and about 45 slopes
at the top portion where the rock mass is signi¢cantly weathered. These slope angles
agree very well with the results obtained for overall maximum safe slope angles in
this study.
The maximum safe slope angles from the kinematic analyses coincide very well
with those for the type I blocks in the block theory. It is not possible to differentiate
the types of blocks using the kinematic analysis. In addition, some conceptual dif-
ferences exist between the kinematic and block theory analyses in estimating the
maximum safe slope angles under gravitational loading. Therefore, a proper com-
parison cannot be made between the two analyses for block types II through V.
For single and double plane sliding, block theory results are much superior to
kinematic results. However, kinematic analysis is important in obtaining maximum
safe slope angles under toppling mode.
It is important to keep in mind that these conclusions are based on the kinematic
and block theory analyses performed using only the major discontinuities. In the
future, kinematic and block theory analyses should be performed incorporating both
major and minor discontinuities. All the analyses conducted in this study limited the
loading on the rock mass to gravitational forces only. Therefore, some equilibrium
stability analyses should be performed incorporating possible water forces,
earthquake forces, etc.

Acknowledgements
Authors would like to thank the Yangtze Water Resources Commission, China for
providing us the discontinuity data.

References
Chen, D. (1986) The geological study of the Three Gorges project in China. In Proc. 5th Int.
IAEG Congress, Buenos Aires, Argentina, pp. 1067^1075.
Goodman, R. E. (1989) Introduction to Rock Mechanics,-2nd Edn, John Wiley and Sons, Inc.,
New York.
Goodman, R. E. and Shi, G.-H. (1985) Block theory and its Application to Rock Engineering,
Prentice-Hall Inc., New Jersey.
Ha, Q. (1993) Rock engineering related to the Three Gorges project, Proc. Int. Symposium on
Rock Engineering Related to the Three-Gorges Project, Yichang, China.
Kulatilake, P. H. S. W. (1998) Software manual for FRACNTWKöa computer package to
model discontinuity geometry in rock masses, University of Arizona, Tucson, 394 pages.
Kulatilake, P. H. S. W., Chen, J., Teng, J., Shufang, X., and Pan, G. (1996) Discontinuity
geometry characterization for the rock mass around a tunnel close to the proposed per-
manent shiplock area of the Three Gorges dam site in China, Int. Jour. of Rock Mech.
and Mining Sciences, 33, 255^277.
Teng, J (1994) Personal Communication.
42 J-G. UM AND P. H. S. W. KULATILAKE

Xuecheng, D. (1993) Stability of the dam foundation, rock mass classi¢cation and
optimization of the foundation line, Proc. Int. Symposium on Rock Engineering Related
to the Three-Gorges Project, Yichang, China.
Wang, X. (1986) Mechanism of weathering for crystalline rock in the Three Gorges and dis-
cussion of physical and mechanical properties of slightly weathered rock mass, J.
Changchun Univ. of Earth Sciences, pp. 93^100.