LESSONS FROM THE FAILURE OF FULL-SCALE MODELS AND RECENT GEOSYNTHETIC-REINFORCED SOIL RETAINING WALLS

1
LESSONS FROM THE FAILURE OF FULL-SCALE MODELS
AND RECENT GEOSYNTHETIC-REINFORCEDSOIL
RETAINING WALLS

• F. Tatsuoka Department of Civil Engineering,
• University of Tokyo
•  
• M. Tateyama Railway Technical Research Institute
•  
• Y. Tamura Integrated Geotechnology Institute
  Ltd
• H. Yamauchi Penta-Ocean Construction Co.

2
ABSTRACT

• 1) Geosynthetic-reinforced soil RWs having a
  full-height rigid facing have been constructed
• - for a total wall length of more than 35 km in
  Japan and
• - as permanent important railway and highway soil
  retaining structures.

3
Geogrid
4

• 2) Staged construction
• the wall is first constructed with a help of
  gabions filled with crushed gravel and

5
????????????????????(??????)?
Sand backfill
6

• 2) Staged construction
• the wall is first constructed with a help of
  gabions filled with crushed gravel and

- then full-height rigid facing is cast-in-place
on the wrapped- around wall.
7

• GRS-RWs having a full-height rigid facing
• constructed by the staged construction procedure
• - now supporting railway and highway embankments
  for a length
• more than 35 km

- have become one of the standard wall
construction procedures for railways,
replacing the conventional procedures and
-no problematic case reported since its
introduction.
8
Locations of major GRS-RWs with a full-height
rigid facing constructed by the staged
construction procedure (as of April 2000).
9
BACKGROUND

• History of elevated railway and highway
  structures in Japan

Gentle slope       could be unstable could be
too deformable and occupies too large
space.  
Some cases
More cost-efficiency Sufficiently stable and
stiff (no piles)
10

• GRS-RWs with a full-height rigid facing

1) A very small interaction between a rigid
facing and deformable backfill during filling-up
and during compacting the backfill - Also, large
deformation of the supporting ground can be
accommodated, without losing the stability of
wall. ? No pile foundations required for
GRS-RWs.

• 2) Full-height rigid facing makes GRS-RWs
• stable rigid (in particular, against load
  applied on the top of facing or the crest of
  wall)
• durable and
• aesthetically acceptable.
• ? equivalent to RC cantilever retaining
  structures.
  (to continue)

11
Different types of retaining walls
12
Conventional type retaining walls cantilever
structures
Reinforced-soil RWs with a full height rigid
facing continuous beams with a number of
supports
13
Conventional type RWs
Load at the base of RW
Wall height H
S
M
Shear load S proportional to H2 Moment M
proportional to H3 Conventional type RWs become
less cost-effective as H becomes larger,
exceeding about 5 m. This is the particular the
case in RWs on slopes in mountain areas.
14
Load at the base of RW
? Load equilibrium along the potential failure
plane that develops in the unreinforced
backfill. Active earth pressure is resisted by
the tensile force developed in the reinforcement
at each height of backfill. The role of
reinforcement is the same with the conventional
wall structure in resisting the earth
pressure. ?but, no large shear load and moment
activated at the base of wall structure,
because the reinforced soil RW is not a
cantilever structure . Then, facing is not
necessary ?
15
Reinforced-soil RWs
?The active zone may fail without adequate
facing.
?Load equilibrium at the facing The earth
pressure acting the back face of the facing
confines the backfill, making the backfill
stable. The earth pressure is resisted by the
connection force of the reinforcement? As a
continuous beam with many supports, a) Very
small shear force and moment working inside the
facing structure, making the facing structure
much simpler than the wall structure of
conventional type RWs. b) Very small shear load
and moment at the bas of the facing structure
The shear load and moment does not increase
proportionally to the wall height (H). ? a pile
file foundation becomes unnecessary.?
16
Role of facing structure in stabilising the
active zone and in developing high tensile force
in the reinforcement
Tensile force in case with facing
Tensile force in case without facing
Development of tensile force in the reinforcement
connected to facing.
17
Goegrid-reinforced soil RW along JR Kobe Line
(1992)
18
Goegrid-reinforced soil RW along JR Kobe Line
(1995)
19
Damaged conventional type RWs (1995) Gravity
type and cantilever RC without a pile foundation
lt?????gt
ltL???gt
20
Bridge abutment (Seibu Line in Tokyo)
21
(No Transcript)
22
Bridge abutment and pier (Sasaguri Line in Kyushu)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
Increasing the seismic stability by preloading
and prestressing with and without a ratchet
system
Without a ratchet system
With a ratchet system
Shaking table test (700gal, 5Hz, 25sec)
35

• (continued)
• 3) Full-height rigid facing makes GRS-RWs
• stable and rigid enough even with relatively
  short reinforcement
• advantageous when reconstructing an existing
  gentle slope to a vertical wall.

36

• A typical latest project at Shinjuku, Tokyo
•  
• - reconstruction of an old bridge and associated
  relocation of two railway tracks for the busiest
  and most important rapid transits in Japan, Chuo
  and Yamanote Lines.

37
Life in Japan (Shinjuku)
38

• - One of the most critical and challenging case
  histories,
• started 1995 and completed in the beginning of
  2000.

39

• Why was GRS-RW having a full-height rigid facing
  selected ?
•  
• 1) high cost-effectiveness
• 2) sufficiently stable and stiff walls to support
  extremely important railways

3) relatively soft subsoil a deep pile
foundation is necessary for conventional
cantilever RC RWs, but not for GRS-RWs
4) a very severe space restraint at the site
large construction plants cannot be used.
GRS-RW needs only small construction plants.
40

• Yodobashi site, Shinjuku, Tokyo

41

• (continued)
• 3) Full-height rigid facing makes GRS-RWs
• stable and rigid enough even with relatively
  short reinforcement
• advantageous when reconstructing an existing
  gentle slope to a vertical wall.

The use of inferior on-site soil, such as sand
including a large amount of fines and even a
nearly saturated clay, becomes possible with a
help of, for example, a composite geosynthetic
having a drainage function of non-woven
geotextile component and a high tensile stiffness
of woven geotextile component.
42
Composite geosynthetic for use in clayey soils
Woven geotextile (tensile reinforcement)
Non-woven geotextile (drainage)
43
Nearly saturated highly weathered tuff Nagano
wall

•  -constructed in 1994 to reconfirm the function
  of full-height rigid facing
• in conjunction of the construction of proto-type
  GRS-RWs for 1993 - 1994.

44
Backfill of nearly saturated clay
45
RESEARCH TO DEVELOP THE GRS-RW SYSTEM-
started early 1980's by small-scale static
loading and shaking table tests in the
laboratory numerical analysis by LEM FEM
and
full-scale failure tests in the field, since
1982.
46
FIELD TEST PROGRAM

• - Three soil types for the backfill
• 1) On-site nearly saturated volcanic ash clay
  (Kanto loam)
• Chiba No. 1 Chiba No. 2
• Kami-Onda Chiba No. 3 and
• JR No. 2.
• 2) On-site nearly saturated highly weathered tuff
  - Nagano
• 3) Sand- JR No. 1.

- Failure tests of wall by 1) natural rain for
a long duration 2) supplying a large amount of
water from the wall crest and 3) vertical
loading at the wall crest.
47
Chiba No. 1 embankment with clay backfill

• - to examine whether vertical stable reinforced
  clay walls can be constructed

constructed in 1982 at Chiba Experiment
Station, the University of Tokyo a non-woven
geotextile (spun-bonded 100 poly-propylene)
- with a drainage function only, without a high
tensile rigidity. nearly vertical flat
wrapped-around wall face.
48
Large deformation already during construction and
also for a long period after construction,
particularly by heavy rainfalls.

• the real serious problem is with wrapped-around
  flat walls.

49
Failure mechanism of Chiba No. 1 embankment

• a) Rain water percolated through the backfill and
  accumulated in the bottom soil layer,
• a reduction in the soil suction and further
  an increase in the positive pore water pressure,
  making the soil very weak

b) The wrapping geotextile could not
effectively restrain the deformation of flat
wall face.
c) Compressive failure behind the wall face
proceeded towards deeper places. The
reinforced soil zone above the bottom soil layer
settled down and displaced outward as a monolith,
creating a shear zone between the reinforced and
unreinforced zones.
50

• Lessons from the failure-1
•  
• 1) It is very important for the wall stability to
  prevent a local soil failure in the soil
  immediately behind the wall face.

2) Flat wrapped-around wall face cannot confine
effectively the soil behind the wall face -
cannot be recommended for important structures.
- If used, wrapped around wall face of each soil
layer should be constructed to be round.

(to continue)
51

• Lessons from the failure-2
•  
• 3) A vertical spacing of 80 cm between geotextile
  layers is too large
• i)   to effectively drain water from each clay
  layer and to maintain a high suction in the soil
  layer and
• ii) to effectively confine the clay backfill, in
  particular immediately behind the wrapped-around
  wall face.

The importance of keeping a sufficiently high
suction in the backfill soil for the stability of
clay wall cannot be over-emphasised.
52
Chiba No. 2 embankment with clay backfill  

• to confirm the lessons from the behaviour of
  Chiba No. 1 and
• to investigate the effects of gabions at the
  wall face on the wall stability.

- constructed in 1984 - slightly larger than
Chiba No. 1 embankment and - the same types of
clay backfill and non-woven geotextile as Chiba
No. 1.
53
- Despite its low stiffness of the non-woven
geotextile reinforcement, the reinforced clay
walls performed very well for the first year.  

• - To bring the walls to failure, 70 m3 of water
  ( 900 mm precipitation) was supplied from a pond
  on the crest for eight days in October 1985.

Pond made on the crest of Chiba No. 2 embankment
and cracks developed by water supply from the
pond.
54

• Lessons from the failure
• 1) Although gabions were filled with the clay
  backfill (i.e., Kanto loam), their use at the
  shoulder of each soil layer was very effective
  for
• a good compaction of backfill and
• confining the soil near the wall face,
  maintaining a high soil strength.

In the later actual construction projects,
gabions are filled with gravel for a much better
functioning, not losing cost-effectiveness.
55

• 2) Major cause for the wall deformation by the
  artificial rainfall test
• - the decrease in the suction and
• - the increase in the positive pore water
  pressure.
•  

? Effects of pore water pressure should be
considered in the stability analysis.
Behaviour of Chiba No. 2 embankment during the
artificial rainfall test.
56

• 3) Three failure modes

Cross-section exposed at its demolishing -
lines 1 2 critical failure surfaces by the
limit equilibrium stability analysis without and
with taking into account the pore pressure in
cracks.

57

• - R L total deformation by the rainfall test.
  
• - Ra deformation only in the last day of the
  rainfall test.
• - K similar data for Kami-Onda embankment.

58

• -Three failure modes

The safety factors for all these three failure
modes should be examined in design.
59

• 4)  Despite the use of so-called very extensible
  reinforcement (i.e., a non-woven geotextile), no
  failure plane and tension cracks in the
  reinforced zones, as with Chiba No. 1 test
  embankment.

60

• 5) Practically no creep deformation of the walls
  after the rainfall test in the second year (1985)
  
• - due to effects of rainfall as preloading.

Behaviour after the artificial rainfall test
until its demolishing.
61
Chiba No. 3 embankment with clay walls

• 1) constructed in 1986 using the same types of
  clay backfill soil and non-woven geotextile as
  before.

62
Chiba No. 3 embankment with clay walls

• 2) to confirm the lessons from the previous three
  tests by comparing the behaviours of
• a) a wrapped-around wall (without gabions)
• b) a discrete concrete panel wall and
• c) a wrapped-around wall (with gabions) covered
  
• with a 8 cm-thick shotcrete layer.

63

• Lessons from the failure
•  
• Different behaviours of the three walls according
  to very different facing rigidities.

1) Reconfirmation clay walls with flat
wrapped-around face without gabions are too
deformable to be used as important structures.
64

• 2) Facing of relatively small discrete panels
• not rigid enough and
• very difficult to compact soil immediately behind
  the facing and to achieve a good wall face
  alignment.

65

• 3) Wall constructed by staged construction
• - behaved well and
• - much better construction
• efficiency than with
• discrete panel facing.

? Even nearly saturated clay can be used as the
backfill when reinforced having a proper drainage
function and a tensile rigidity, together with a
proper rigid facing.
But, 1) a 8 cm-thick shotcrete layer was not
rigid enough to keep the wall face deformation
small, and 2) shotcrete facing may not be
aesthetically acceptable.
66
JR No. 2 embankment with clay backfill

• -constructed at Japan Railway (JR) Technical
  Research Institute in the beginning of 1988.

-to examine the stability of GRS-RWs having a
full-height rigid facing nearly the same wall
structure as the final one that was later used in
the actual projects.
67

• Clay backfill (Kanto loam) wi 120 - 130 Sr
  90 and ?d 0.55 - 0.60 g/cm3.
• Three types of reinforcement
• section a) a non-woven geotextile, as used for
  the other
• embankments
• section b) a grid sandwiched between two gravel
  drainage layers
• and
• section c) a composite consisting of
  non-woven/woven
• geotextiles.

68

• - A very good and similar performance of the
  three sections for a long duration, reconfirming
  thatthe facing type could be much more important
  than the stiffness of reinforcement for the
  stability of reinforced soil retaining wall.

69
Nearly saturated highly weathered tuff Nagano
wall

•  -constructed in 1994 to reconfirm the function
  of full-height rigid facing
• in conjunction of the construction of proto-type
  GRS-RWs for 1993 - 1994.

70
Nearly saturated highly weathered tuff Nagano
wall

• a) a complete wall height of 2 m for a length of
  2 km, supporting a yard for Shinkansen (bullet
  train)
• b) the first actual clay wall using a nearly
  saturated soft clay as a railway structure in
  Japan,
• c) constructed on a thick very soft clay deposit
• d) a large ground settlement of about 1 m by
  preloading before casting-in-place a rigid
  facing and
• e) no pile foundation.

71

• JR No. 1 test embankment with sand backfill

constructed for a period of 1987 - 1988 - to
examine the stability of GRW-RWs with sand
backfill having a full-height rigid facing.
Grid tensile strength 2.8 tonf/m. Wall
structure very similar to the one that was
subsequently employed for actual projects.
72

• JR No. 1 test embankment with sand backfill

Development of staged construction procedure
73

• Two types of facing
• segment h a discrete panel facing fixed to
  gabions filled with gravel.

the other segments a full-height rigid facing.
74

• Long-term behaviours for about two years
•  
• a) Segment h, having a discrete panel facing
• - much larger deformation than that the other
  walls.

b) the other segments, having a full-height rigid
facing -practically nil deformation.
75

• Loading test of No. 1 to failure
• after very stable behaviour for about two years

Loading method for JR No. 1 sand embankment
76

• FAILURE AND LESSONS - JR No. 1 embankment
•  

1) Segment h, having a discrete panel facing
- most deformable and weakest, not
relevant to permanent important structures.
77

• 2) f (L 1.5m) vs. d (L 2.0 m).
•  

Segment f (L 1.5m) - less stable than segment
d but stable enough against ordinary design
loads.
- The current design method the minimum
allowable reinforcement length is equal to or
longer than the smaller value of a) 1.5 m
and b) 35 of the wall height, on the
premise that the wall stability is examined by a
proper stability analysis.
78

• 3) Failure at the construction joint in the
  unreinforced facing controlled the yielding of
  the test wall segments f and d (No. 1
  embankment).
•  

The facing used for prototype GRS-RWs is lightly
steel-reinforced to withstand the design earth
pressure, which is equal to the active earth
pressure when the backfill soil is not
reinforced.
79

• 4) Two-wedge failure mode in wall segment h

Even with so-called extensible reinforcement, the
development of failure plane in a reinforced zone
is very difficult !
Shear zone observed in segment h having a
discrete panel facing.
80

• Loading test to failure of JR No. 2 clay wall
• after very stable behaviour for about two years

81
1) Stronger when loaded on the crest close to the
wall face due to the effects of reinforcement.
82

• 2) When properly reinforced, a clay wall is not
  very weak,
• compararable with a sand wall !

83

• 3) Gabions as a buffer for the relative
  settlement between the rigid facing and the
  backfill soil

- preventing damage to the connection between the
facing and the reinforcement.
facing
b)
Cross-section after loading test.
84

• 4) Any clear failure plane in the reinforced zone
  !

facing
a)
Cross-section after loading test.
85
CONCLUDING REMARK-1 Wrapped-around walls are
generally too deformable, particularly when the
wall face is finished flat, to be used as
permanent important structures allowing a limited
amount of deformation.  
86
CONCLUDING REMARK-1 Wrapped-around walls are
generally too deformable, particularly when the
wall face is finished flat, to be used as
permanent important structures allowing a limited
amount of deformation.  
A rigid facing, in particular a full-height
continuous rigid facing to which reinforcements
are fixed, helps in increasing the stability of
wall and in decreasing the deformation of wall.
87

• CONCLUDING REMARK-2
•  
• The construction of sufficiently stable and rigid
  clay walls as permanent important structures is
  quite feasible by reinforcing the backfill with a
  proper composite geotextile having a sufficiently
  high drainage function and a tensile rigidity and
  by using a full-height rigid facing.
•  

88

• CONCLUDING REMARK-2
•  
• The construction of sufficiently stable and rigid
  clay walls as permanent important structures is
  quite feasible by reinforcing the backfill with a
  proper composite geotextile having a sufficiently
  high drainage function and a tensile rigidity and
  by using a full-height rigid facing.
•  

In the design of GRS-RWs having clay backfill,
due consideration of drainage and consolidation
of clay soil layers between geosynthetic
reinforcement layers is essential.
89

• CONCLUDING REMARK-3
•  
• A number of prototype GRS-RWs have been
  constructed for a total wall length being about
  35 km by the staged procedure for the last decade
  in Japan.
•  

It is significant that so far any problematic
case has been reported.
90
Thank you very much for your kind attentions !