Rehabilitation of a Failing Anchored Retaining Wall - Nicholson ...

Nicholson Construction Company
12 McClane Street
Cuddy, PA 15031
Telephone: 412-221-4500
Facsimile: 412-221-3127
Rehabilitation of a Failing Anchored Retaining Wall
by
Thomas Richards
Nicholson Construction Company, Cuddy, Pennsylvania
Daniel Thome
Nicholson Construction Company, Kalamazoo, Michigan
Presented at:
Institution of Civil Engineers
International Conference on Ground Anchorages and Anchored Structures in Service 2007
London, United Kingdom
November 26-27, 2007
07-04-159

Richards/Thome, Page 1
Rehabilitation of a Failing Anchored Retaining Wall
Thomas Richards, Nicholson Construction Company, Cuddy, PA USA
Daniel Thome, Nicholson Construction Company, Kalamazoo, MI USA
Abstract
Significant movement to an existing diaphragm (slurry) wall on ‘O’ Street in Southeast
Washington, D.C. was caused by failure of some anchors. Potential causes of failure are
presented. Rehabilitation and stabilization for the failed wall was performed by demolishing
the failed panels, drilling horizontal weep holes, constructing a new precast lagging and
soldier pile wall in front of the existing structure, and improving site drainage system.
Background
In the late 1970s the District of Columbia decided that permanent stabilization of the slope
was required due to the continuous slope movements toward ‘O’ Street. In 1978 an anchored
slurry wall was constructed between ‘O’ Street and Highwood Drive to retain the earth from
further movements.
The anchored slurry wall consisted of fifty, 6.7 meter wide reinforced concrete panels
designed to be embedded into a hard clay stratum. High strength threaded bar anchors were
installed through the slurry wall and designed to be bonded into the hard clay. During
construction the bearing plates to the ground anchors were reported to have been visually
showing failure after they were stressed. Eventually the initial bearing plates were replaced
with larger ones soon after construction. Corrosion protection was a tar like coating applied
to the bearing plates and beveled nuts, and anchor heads after they were installed and stressed.
In the winter of 1995 excessive ground movements were noted in approximately 46 meter of
the slurry wall. Ground movements were recorded on this particular section of the wall and
continued until the spring of 1996 when eventually this section failed. Maximum deflections
in the order of 4 meters downslope were recorded. Through time the ground anchors became
overloaded and failed causing a global stability failure with a distinct slide plane reaching
completely under the toe of the slurry wall.

Richards/Thome, Page 2
Photo 1 Overview Distressed Section of Wall
Photo 2 Close-up of Leaning Panels (wall in distance is vertical)
An investigation (Thomas L. Brown Associates, P.C, 1999) of the wall determined that the
global failure was caused by three contributing factors:
1. Improper drainage from behind the wall is believed to be the major cause of the stability
failure due to the additional hydrostatic pressure acting on the slurry wall
2. Zones of highly plastic slickensided clay below the toe of the wall

Richards/Thome, Page 3
3. Lack of corrosion protection on the initial thread bars used as tieback supports.
Conclusions of the study suggested that while each defect in the wall is believed to be a
contributing factor in the failure, the combination of all three factors acting simultaneously
probably initiated the global failure. Improper drainage from behind the wall increased the
earth pressures acting on the wall leading to overload the tiebacks, failing in part due to the
lack of sufficient corrosion protection around the ground anchors. Finally, subsurface
investigations concluded that the toe of the slurry wall was actually embedded in a stratum of
slickensided clay, which created a less resistant path for the failure plane.
Condition of Old Anchors
Bar anchors were installed in 1978. As noted above, the bearing plates and nuts reportedly
exhibited problems during testing. By 1999, several of the anchors had failed causing the
wall to lean significantly. Detailed condition assessment of the anchors was not a component
of our work, since the rehabilitation did not rely on any of the existing anchors.
As part of demolition of leaning wall panels, several anchors were exhumed. The breaks were
typically within the wall or just behind the wall as shown in Photos 3 and 4. The pipe was a
typical anchor sleeve in the diaphragm wall.
Photo 3 Breaks in Anchors behind Wall
Photo 4 Breaks in Anchors behind Wall
Grouting of the sleeves seemed to have some deficiencies as shown in Photos 3 and 4 and also
several of the other anchors which had seepage and grass growing around the bearing plates.
It is also unclear how access for secondary grouting was provided.
The broken bar had pitting corrosion and had been protected by plastic sheath as shown in
Photo 5. It did not appear that grease had been used or remained inside the plastic sheath.
Water could then get to the bars through the incomplete secondary grouting and then travel
down the bar. However, the bars would then have been expected to fail near the top. This
suggests that shear loads may have also been present at the back face of wall to ground
interface due to possible filling behind the wall or lateral wall movement allowing downward
soil movement behind the wall.

Richards/Thome, Page 4
Photo 5 Close-up of Failed Bar
Several other anchors were found to have broken and popped out of the wall as shown in
Photos 6 to 9.
Photo 6 Popped Anchor note gap between
nut and washer and between bearing plates
Photo 7 Missing Anchor and Anchorage

Richards/Thome, Page 5
Photo 8 Popped Anchor – note gap between
washer and bearing plate and between
bearing plates
Photo 9 Popped Anchor
In general, the tar performed well in protecting the bearing plates and nuts, which had less
corrosion than the photos above.
Most Aggressive Design Section near Failures
The area of the wall with most of the anchor failures was judged by the authors to be the most
aggressive design section. Photos 1 and 10 show that the most distressed section of wall was
just beyond a transition from two rows of anchors to one row. The elevation of the ground
near houses behind the wall was similar and a significant slope (4.5m high at 2H:1V) existed
behind the failed section. This slope and change in top of wall elevation also created a
significant water runoff area focused on the failed area with one catch basin inlet. The catch
basin inlets were poorly maintained and many were clogged with debris frequently leaves and
vines. After failure of anchors in the one row section (right of Photo 10), prying action of the
continuous cap beam then would have lead to overstressing of two row section.
Photo 10 Existing Wall Elevation ( left of Photo 1)
Probable Cause of Failures
The failure of the anchors is judged by the authors to have been caused by a combination of:
1. Buildup of water pressure due to clogging of weepholes in the existing wall

Richards/Thome, Page 6
increasing the load on the anchors
2. Most aggressive design section with a backslope surcharge
3. Movements of the wall in a landslide prone area with questionable toe support
4. Corrosion resulting from 1 and 3
Stabilization of the Existing Wall
The section presents the construction of the new wall.
Demolition of Leaning Panels
For safety and access to build a smooth new wall line, approximately 45 meter length of the
failed slurry wall was demolished down to the ground line behind the wall up to 3 meters
below the top of the wall. The ground line was below the top of wall due to the considerable
lateral wall movement.
Weepholes
Nine weepholes were installed through the downhill face of each panel to relieve water
pressures built up behind the existing slurry wall. Each weephole consisted of a 38 mm
diameter well screen covered with a filter fabric sock to prevent any loss of ground from
occurring behind the wall. These holes were drilled through the existing slurry wall and
continued approximately 0.5 meters into the soil in order to promote a proper drainage path.
The original weepholes were simply plastic pipe through the wall apparently with filter
protection.
Test Anchor Program
The anchors were founded in hard clay below the slickensided zones. A total of eight test
anchors each with six strands were installed with differing drilling techniques and casing size.
Post grout tubes were also installed along the length of the additional anchors with rubber
sleeves covering holes with in the bond zone. We found that the clay was severely softened
using water or significant amounts for dust suppression. We chose air as a flush fluid to
minimize mess in a residential sloping site and minimal water for dust suppression.
New Wall Construction
A new steel soldier pile and precast lagging wall was constructed in front of the slurry wall.
Precast wales were used to connect the anchors to the soldier piles.
The detail section is shown Figure 1.

Richards/Thome, Page 7
Figure 1 Rehabilitation Typical Design Section
Soldier piles were placed in 0.9 meter predrilled holes. To reduce risk of any slurry wall
movement some shafts were skipped and then drilled on another shift. Slurry wall panels
were monitored with a total station approximately every ten minutes during drilled shaft
installation by shooting prisms mounted to the slurry wall. Readings continued on the wall
every half an hour until the soldier piles were placed in the drilled shafts and filled with
structural concrete.
Due to the precast lagging and especially the precast wales, accurate soldier beam placement
was required. 10 meter long templates were made that would support the beams in the drilled
shafts prior to the initial set of the structural concrete.
Prior to setting the concrete waler, existing anchor locations on the slurry wall were compared
to proposed anchors through the new wall. If any of the new anchors were expected to hit an
existing anchor bearing plate, the waler would be slid to one side if clearance of the bearing
plate could be accomplished. As a last resort a portion of the bearing plate would be burned

Richards/Thome, Page 8
off to clear the way for the path of the new ground anchor. Nonshrink grout was poured in
the space between the soldier pile and waler.
Concrete was placed between the precast wale and the slurry wall to:
• reduce deflection in the new wall during anchor stressing operations by
transferring some of the load in the new wall to the existing slurry wall and
• provide a conduit so that drill cutting would not contaminate the gravel drainage
layer behind the wall.
PVC sleeves were placed prior to concrete to connect the two gravel layers above and below
the waler concrete.
Ground Anchor Drilling and Installation
Photo 11 Drilling of New Anchors
A total of 262 anchors were installed with the following details:
• Three to six strands per tendon
• Strand – ASTM A416 15.2mm diameter 261 kN ultimate strength
• Full length corrugated HDPE sheath - 90 mm OD, 76 mm ID, 1.5mm thick
• Free lengths of 16.8 m with grease and extruded sheath on individual strands
Number of Strands Working Load (kN) Bond Length (m)
3 605 10.7
4 818 12.2
6 1063 15.2
Duplex drilling with air was used for the majority of the taller end and through the failed
section of the wall. For the remainder of the project, duplex drilling was performed for the
upper 10 to 16 meters through the softer saturated clay and sand lenses and then the hard clay
was drilled open hole. Surface casing was longer on the eastern side of the site because of the
deeper saturated sandy silt found below the wall.

Richards/Thome, Page 9
For anchors installed in an open hole, great care was taken to clean out the hole prior to
anchor installation. PVC centralizers, 127 mm in diameter, were used to keep the corrugated
corrosion protection sheath from dragging through the soil and to concentrate the anchor in
the center of the hole when installed. These centralizers were placed on the corrugated plastic
sheath at 1.5 meter spacings throughout the bond zone. Results showed clean holes with good
air return when casing went beyond the problem zone.
Anchors (six strand and about half of the four strand) in the fail zone and tall wall sections
were post grouted with approximately six bags per tube. Post grouting ceased after operations
went out of the failed portion of the wall. Ground anchors out of the failed zone were only
post grouted when the stressing criteria were not met.
The grout mix was neat cement grout with Type I (ASTM C150) cement. The water to
cement ratio was 0.45 for initial grouting and 0.67 for post grouting.
Each anchor was tested to 133% of design load and accepted according to the Post-
Tensioning Institute, “Recommendations for Prestressed Rock and Soil Anchors” (1996).
Secondary Grouting and Closure Panel Installation
Anchor tails from stressing were removed once the anchor data was proved to conform to the
stressing criteria. Approximately two bags of grout were then placed in each anchor head to
flush any remaining water existing in the hole. This was done to prevent any water from
coming in contact with the anchor in the future. If particular anchors were found to be
seeping water they were pressure grouted to remove the excess water and seal the anchor head
with grout.
A 50 mm thick closure panel, approximately 2 feet high by 5 feet long, was placed over the
two anchors in the recessed portion of the waler. The closure panel was temporarily braced
from the front while grout was pumped from holes placed on top of the waler. Once the grout
cured the closure panel was held in place by a metal loop previously precast into the panel and
grouted in during closure panel installation.
Photo 12 Precast Cover Panel Installation and Grouting
Drainage
Water from in between the existing slurry wall and the new precast wall had to be diverted
once it went through the gravel backfill. A wall “French” drain was installed approximately
18 inches in front of the existing precast lagging wall. Excavation for the drain occurred

directly in front of the wrapped filter fabric set prior to precast panel installation.
Richards/Thome, Page 10
Downhill of the wall, surface drainage swales and numerous new french drains were installed
to increase the probability of intersecting the groundwater seepage running from underneath
the precast lagging retaining wall.
Conclusions
Failure of some permanent bar anchors led to an extensive rehabilitation of the wall. The
failure is considered to have been caused by a combination of
1. Buildup of water pressure due to clogging of weepholes in the existing wall
increasing the load on the anchors
2. Most aggressive design section with a backslope surcharge
3. Movements of the wall in a landslide prone area with questionable toe support
4. Corrosion resulting from 1 and 3
The finished wall provided much improved stability and more pleasing aesthetics.
Photo 13 Finished Wall
References
Thomas L. Brown Associates, P.C., “Pre-Design Studies ‘O’ Street Retaining Wall and Soil
Movements”, Prepared for Department of Public Works District of Columbia, February 1999
Thome, Richards, Wise, “ ‘O’ Street Anchored Precast Lagging and Soldier Pile Retaining
Wall” Contemporary Solutions to Land Mass Stabilization, IX Great Lakes Geotechnical
Geoenvironmental Conference, Dayton, Ohio, May 11, 2001