See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/274665908

Stability Inspection Method for Existing Retaining Walls

Article  in  Quarterly Report of RTRI · August 2013

DOI: 10.2219/rtriqr.54.159

CITATIONS READS

3 697

5 authors, including:

Masahiro Shinoda Keita Abe

National Defense Academy of Japan Railway Technical Research Institute
108 PUBLICATIONS   388 CITATIONS    53 PUBLICATIONS   376 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Seepage failure View project

Retaining wall View project

All content following this page was uploaded by Masahiro Shinoda on 14 February 2017.

The user has requested enhancement of the downloaded file.

 PAPER

Stability Inspection Method for Existing Retaining Walls

Masahiro SHINODA, Dr. Eng. Susumu NAKAJIMA, Dr. Eng. Keita ABE, Dr. Eng.
Senior Researcher, Assistant Researcher, Assistant Researcher,
Foundation and Geotechnical Engineering Laboratory, Structures Technology Division

Toshiaki EHARA Yuki KUBOTA

Researcher, Researcher,
Foundation and Geotechnical Engineering Laboratory, Structures Technology Division

Visual inspection methods are currently used on existing retaining structures. Conse-
quently it is difficult to evaluate a structural health of the retaining walls quantitatively. In
this study, a series of model tests on retaining walls were carried out to develop an inspection
method for quantitatively evaluating existing railway retaining walls. It has been found from
the series of model tests that vibration tests are an effective means to determine the vibration
characteristics of retaining walls, which reflect their structural health. Test results indicated
that small scale vibration tests could be applied to evaluation the structural health of existing
retaining structures.

Keywords: inspection, retaining wall, vibration test

1. Introduction 2. Current state of railway retaining structures

The high number of old railway structures in Japan 2.1 Maintenance standards in Japan
emphasizes the importance of having proper maintenance
methodologies. Proper management of railway structures There are approximately 30 thousand kilometers of
requires early detection of deformation in structures. Once railway lines in Japan, operated by several railway organi-
deformations are detected, continuous observation and zations (seven Japan Railway companies, over 100 private
retrofitting work are also important. Methods for quanti- railway companies and several local governments). Japa-
tatively assessing the structural health of railway bridge nese railway organizations maintain their structures in
substructures in Japan have already been developed.
START
In the case of retaining walls however, visual inspec-
tion is still used to evaluate the structural health of ex- (Interval ; within 2 years)
isting retaining walls because a quantitative inspection Condition rating “B,C,S”
General Inspection
method has not yet been developed. And yet, a quantitative
evaluation method is urgently needed, given the subjective Condition rating “A”
Condition rating “B,C”
nature of purely visual inspections. Individual Inspection
Given this background, the present study aims to de-
velop a quantitative inspection method for establishing the Condition rating “B” Condition rating “A”
structural health of existing retaining walls. This paper Countermeasures/Retrofitting works
begins with a brief introduction of the current state of ex- /Monitoring
isting retaining structures. Secondly it gives a brief outline
of the small scale exciter, which has been newly developed
Records
to evaluate the structural health of retaining walls. Lastly,
results are discussed of model tests conducted to determine Fig. 1 Maintenance procedure according to the Stan-
the applicability of the small scale exciter. dards for Maintenance in Japan

Table 1 Relationships between soundness and structural condition)

Condition rating Structure state
State that threatens operational safety, safety of passengers, public
A safety, guarantee of regular train operation, or deterioration that
might cause this situation.
B Deterioration that might result in A rate condition in the future.
C Slight deterioration.
S Good condition.

QR of RTRI, Vol. 54, No. 3, Aug. 2013 159

 safe condition by referring to the Japanese Maintenance U-shaped wall
Code of Maintenance Standards in Japan [1]. General pro- Cantilever type wall 1.8%
5.2% Counterfort wall
cedures for structural maintenance in the Maintenance 0.1% Type of wall Site No.
Gravity type wall
standards and relationships between the condition rating 7.9% Masonry/Ashlar block 6,106
and the structure state are shown in Fig.1 and Table 1. Leaning type 688
Leaning type wall
As indicated in Fig. 1,“General Inspections”are con- 8.6% Gravity type 629
ducted on all railway structures at maximum two-year in- Cantilever type 413
tervals, and are mainly visual.“Individual Inspections”are Masonry wall U-shaped 145
Ashlar block wall
made however on specific structures in which more severe 76.4% Counterfort wall 8
deterioration has been detected during the General Inspec- Total 7,989
tion by means of detailed visual surveys or using measure-
ment equipment. As mentioned above in the Introduction, Fig. 3 Number of retaining walls by type
this study aims to develop a methodology which can be
U-shaped wall
used to quantitatively evaluate the structural health of re- 5.3% Counterfort wall
taining walls as an alternative to detailed visual surveys. Gravity type wall
0.2%
6.2% Type of wall Number Length(km)
Leaning type 409 41.1
2.2 Survey of current state of Japanese railway re- Cantilever type Masonry/Ashlar block 802 40.6
taining walls wall
12.1% Leaning type Cantilevet type wall 249 13.0
wall
38.3% Gravity type wall 85 6.7
A preliminary survey on current state of Japanese Masonry
U-shaped wall 105 5.7
Ashlar block wall
railway retaining structures was conducted using the 37.8% Counterfort wall 7 0.2
“Structural Management Supporting system (SMS)”[2]. Total 1,657 107.3
The SMS is the database system which contains compre-
hensive information about all types of railway structure, Fig. 4 Construction length of retaining wall by type
such as RC structures, steel structures, foundations, abut-
ments, retaining walls, tunnels and slopes etc. The SMS inputting data about their own structures.
was developed in association with 14 Japanese railway or- In the preliminary survey, information about typical
ganizations and the Railway Technical Research Institute. types of retaining wall in Japan (see Fig. 2) was extracted
Currently, 27 railway organizations utilize the SMS by from the SMS database. In total, data for 7,989 sites was
extracted.
Masonry (or Ashlar) type retainig wall Leaning type retaining wall
Figure 3 shows results of the survey which arranges
the data by type of retaining wall. The numbers of masonry
and ashlar block retaining walls exceeded 6,000 and ac-
counted for 76.4 % of the total data. The number of leaning
type retaining wall was the second largest, and accounted
for 8.6 % of the total number of sites. Valid data contain-
ing the construction length was extracted from among the
total 7,989 sites, since construction length is also an impor-
tant parameter in addition to the number of sites. Figure
4 shows retaining walls and the construction lengths of
Cantilever type retaining wall Gravity type retaining wall
retaining wall by type, which were obtained by averaging
valid data from 1,657 sites.
The construction length of leaning type retaining walls
stands first among all types of retaining walls and ac-
counted for 38.3 % of the valid data. The percentage of the
masonry and ashlar block retaining wall reached 37.8 %. It
was found from the above survey that the leaning type and
masonry or ashlar block retaining wall occupied 76.1 % of
the total construction length underscoring the importance
of managing this type of structures. It was also found from
the survey that the 85 % of leaning type retaining walls
U-saped retaining wall Counterfort wall
were lower than 4 m in height.

2.3 Typical deformation of railway retaining walls

Deformation of railway retaining structures can be di-

Fig. 2 Types of retaining wall
vided into two groups: deformation due to destabilization,
and deformation due to deterioration. Typical deformation
of railway retaining structures is schematically illustrated
in Fig. 5.
Settlement, inclination, swelling, difference in level
and difference at construction joints due to external thrust

160 QR of RTRI, Vol. 54, No. 3, Aug. 2013

can be categorized as deformation due to the destabiliza- +
tion. Exfoliation of concrete and clogging of the drainage
facilities are categorized as deformation due to deteriora- Girder
tion.
Cyclic load due to the train passing, increase of earth
pressure due to the additional construction of the em- Iron ball
bankment, increase of dynamic earth pressure due to the Velocity
earthquake, increase of water pressure due to the change sensor
of the water level in backfill soil are deemed to be sources
of external thrust, which may lead to deformation due to
destabilization. Deterioration of the structure is attributed Bridge pier
to cyclic change in thermal or humid conditions during the
structure’s serviceable life.
Deformation due to the destabilization may lead to
secondary deformation such as backfill loosening or bear-
ing capacity failure. Therefore, early detection and retro-
fitting work to prevent deformation due to destabilization
are particularly important, while no non-destructive tests
to detect such problems have yet been developed. Follow-
ing on, this study attempts to propose a method for non-
destructive evaluation of existing retaining walls.

Difference at
Exfoliation of concrete Ground
Dislocation at
construction joint construction joint
Fig. 6 Outline of percussion test

4
Inclination a)
Velocity (kine)

2
Cracking
Drainage pipe 0
Construction joint
-2
Clogging of drainage facilities
-4
Settlement 0 1 2 3 4
Time (s)
Fig. 5 Typical deformation of retaining wall 0.03
Amplitude (kine * s)

b) Resonance (7.69 Hz)

3. Development of small scale exciter
0.02
3.1 Percussion tests
0.01
Non-destructive testing has been carried out in Japan
on bridge substructures, in the form of percussion tests. In
the past, of the health of railway bridge substructures was 0.00
evaluated by dynamic settlement, frequency and amplifica- 0 5 10 15 20 25
tion of vibration during the passing of trains in addition to Frequency (Hz)
the static measurement of inclination or settlement of the 360
bridge piers. However, this method is insufficient to evalu- c) Resonance (7.69 Hz)
Phase (degree)

ate the performance of bridge substructures quantitatively 270

because measurements made during the passing of trains
mentioned above, depend heavily on the weight and pass- 180
ing speed of the trains.
In order to reduce the above-mentioned problem, a dif- 90
ferent type of percussion test was proposed by Nishimura
et al (1988) [3]. In the latter, the natural frequency of the
0
bridge pier is measured with high accuracy and is used for 0 5 10 15 20 25
evaluating the structural health of the pier. This method is Frequency (Hz)
based on the knowledge that the natural frequency of the
bridge substructure decreases with damage and increases Fig. 7 Typical result of percussion test

QR of RTRI, Vol. 54, No. 3, Aug. 2013 161

with reinforcement.
The natural frequency of bridge piers is found by carry- 200

Acceleration (gal)
ing out a spectrum analysis using measured free vibration, a)
which is recorded by velocity sensors. Free vibration is in- 100
duced by striking the top of the piers using an iron ball, as
schematically illustrated in Fig. 6.
0
A time history of the velocity obtained from the per- -100
cussion test is shown in Fig. 7a). The natural frequency of
the structures can be evaluated based on the Fourier and -200
phase spectrum as shown in Fig. 7b) and 7c), respectively. 0 20 40 60 80 100
The maximum amplitude for the frequency 7.69 Hz is thus Time (s)
obtained as the resonance in accordance with the result of 200

Acceleration (gal)
the phase spectrum in which the resonance occurred in the b)
phase of 180 degrees. In practice, the actual performance of 100
a bridge pier can be evaluated by comparing the measured
natural frequency with that obtained immediately after
0
construction or with the criterion for potential natural fre- -100
quency. Potential natural frequency is the experimentally-
based value proposed by the Railway Technical Research -200
Institute which can be applied to structures in a particular 0 2 4 6
site where the natural frequency was not recorded immedi- Time (s)
ately after the construction. 0.010
Amplitude (gal.s)
c)
3.2 Development of the small scale exciter

As mentioned above, percussion tests have been con-

0.005
ducted on bridge substructures. Applying the same meth-
ods to determine the structural health of retaining walls
poses the following problems: 0.000
1. An input load with a high frequency component can- 0 20 40 60 80 100
not be applied because the input is applied by strik- Frequency (Hz)
ing the top of the pier.
2. The weight of the iron ball (30 kg in general) makes Fig. 9 Example of input from small scale exciter
it difficult to handle.
3. Amplitude of the input load is difficult to measure Table 2 Performance of small scale exciter
quantitatively. Rated sine force (N) 98
4. Amplitude of the input load cannot be kept constant
because the load is applied by different inspector Max. Displacement (mm) 80
each time.
Max. Velocity (cm/s) 74.8
The small scale exciter developed for this study is Max. acceleration (m/s2) 14
schematically illustrated in Fig. 8. A permanent magnet is
adopted for the magnetic circuit mainly because of the por- Frequency range (Hz) 3-200 (at Max. Acc.)
tability despite the resulting difficulty to apply large am-
System power AC100V
plitude and high frequency component as the input. In the
prototype exciter, the vibration force is induced by sending Weight (N) 147
electronic current to the driving coil in the magnetic field.
Dimensions
Neodymium permanent magnet
Driving coil 　Length mm 358
　Width mm 336
　Height mm 194
Direction of force Magnetic Flux density

As schematically illustrated in Fig. 8, an input force vector

B : Magnetic flux density [T]
F : Force [N] I : Magnetic current [A]
Magnetic circuit
Fleming’s left hand rule
Fig. 8 Outline of prototype small scale exciter
is determined by the direction of electric current and mag-
netic flux density based on the Fleming’s left hand rule.
The amplitude of force can be evaluated in eq. (1).
F = BLI (1)
where B is the value of magnetic flux density, L is the
length of the cable in the magnetic flux, and I is the mag-
netic current. In this study, the following parameters were

162 QR of RTRI, Vol. 54, No. 3, Aug. 2013

used: B = 0.49 T, L = 26.7 m and I = 7.5 A. Performance of Displacement transducer 90 Two component loadcell
the developed small scale exciter is summarized in Table 2. 50 +
Examples of the acceleration time history and its Fou- 100
200
rier spectrum induced by the prototype small scale exciter +
are shown in Fig. 9 a), b) and c). Sinusoidal sweep excita- +
tion is usually adopted as the input acceleration to apply
300
the same amplitude of force to the target structure over all +
the frequency components as shown in Fig. 9 b). Applicabil- +
300

ity of the prototype small scale exciter is discussed later in + Accelerometer

the paper with the introduction of the static loading tests +

1000

.35
conducted using a model retaining wall. 300

.3
1:0

1:0
4. Application of the small scale exciter to evaluate 100
+
500
the structural health of a retaining wall

4.1 Outline of the model test 200

As discussed in the introduction to the percussion test,

vibration characteristics of the structure can be used as +
140
Two component loadcell
an index to evaluate the structural health of the retaining
wall. Therefore, a set of the loading tests on the retaining Fig. 11 Cross section of leaning type retaining wall mod-
wall were conducted to investigate the correlation between el (units in mm)
vibration characteristics and displacement of the model re- 100
taining wall.
The test apparatus used in the model test is schemati- 80
80
Horizontal displacement

cally illustrated in Fig. 10. The test apparatus consisted of

of soil chamber (mm)

Displacement
変位

a steel strut, shear soil chamber, hydraulic loading jacks.

The soil chamber was 3 m high with a width and length of 60
2 m.
Models of leaning and masonry type retaining walls 40 40
were placed on a horizontal subsoil consisting of dense air-
dried silica sand. The total height of the model retaining 20 20
wall was 1 m while the embedded depth was 0.2 m. A back- 10
5
fill layer was also constituted with air-dried silica sand. 0 0
The relative densities of the subsoil and backfill layer were
90 and 80 %, respectively. 0 2 4 6 8 10
A cross section of the model leaning type retaining
Loading step
wall is shown in Fig. 11. Two component load cells were in-
stalled on the wall facing and at the bottom of the footing. Fig. 12 Loading process
The outer size of the masonry type retaining wall model
was the same as that of the leaning type retaining wall Horizontal displacement was applied to the shear soil
model, while the wall facing was made by ashlars. chamber using three horizontal hydraulic loading jacks
as shown in Fig. 12. The displacement amplitudes of the
middle and top loading jacks were controlled so that the
soil chamber showed the simple shear mode of deforma-
Soil chamber Upper jack tion as schematically illustrated in the Fig. 12. Vibration
tests using the developed small scale exciter were carried
out during each of the loading and unloading stages. High-
Middle jack
sensitivity accelerometers were attached to the wall to
35

measure changes in vibration characteristics as displace-

2400

1:0.

3
1000

ment of the wall facing increased.

1:0.

Bottom jack
Silica sand 4.2 Test results
Dr=80%
200

Changes in the transfer function of the amplitude of

Silica sand Loading the leaning and masonry type retaining wall models are
635

Dr=90% platen shown in Fig. 13 and 14. The transfer functions corre-
sponding to the initial state and those at a displacement
amplitude of 80 mm were compared in these figures. In the
initial state, a clear peak can be observed around 40 Hz
Fig. 10 Layout of test apparatus both in leaning and masonry type retaining walls.

QR of RTRI, Vol. 54, No. 3, Aug. 2013 163

 0.20
Transfer function (amplitude)

Horizontal disp.（mm）
120
120mm 36.6 Hz 41.2 Hz
90 80mm
Initial Large spectrum area Threshold value
0.15 60
40mm
state

Amplitude
30
10mm
→Deformed retaing wall
0
0mm Small spectrum area
0 2 4 6 8 10 12 14
0.10 Loading step
→Without any deterioration
80 mm loaded
0.05

0.00
10 20 30 40 50 Frequency (Hz)
Frequency (Hz)
Fig. 15 Schematic diagram showing differences in the
Fig. 13 Change in vibration characteristics of leaning Fourier spectrum of amplitude
type retaining wall model
Fourier spectrum, hereafter referred to as the spectrum
0.60 area, is used as an index to evaluate the structural health
of the retaining structure. In the case deterioration, the
Transfer function (amplitude)

Horizontal disp.（mm）

120mm
120
90 80mm
Just after opening value of spectrum area is larger than when the retaining
60 40mm
of masonry joint wall is structurally healthy. Before applying the proposed
30
0.40 0
10mm
Initial state method in practice, a threshold value, indicating the safety
0mm
of the retaining wall would have to be determined. The va-
40.0 Hz
0 2 4 6 8 10 12 14
Loading step
lidity of the proposed method for inspecting existing struc-
tures has been confirmed by Nakajima et al (2012) [4].
0.20 18.8 Hz A greater number of site tests remain to be performed,
in order to meet the shortfall in spectrum area data which
Loaded state is needed to set a threshold value. Consequently, until
enough data is collected, assessing the structural health of
a retaining wall must be achieved by comparing the spec-
0.00 trum areas of a healthy retaining wall with that of a dete-
10 20 30 40 50
Frequency (Hz) riorated retaining wall.

Fig. 14 Change in vibration characteristics of masonry

type retaining wall model 5. Summary

In the case of the leaning type retaining wall, the value The discussion in this paper can be summarized as fol-
of the peak amplitude frequency decreased as displacement lows:
grew and the peak amplitude value for a displacement 1. The management of leaning type and masonry type
of 80 mm was larger than that in the initial state. In the retaining walls was identified as being of primary
masonry type retaining wall model however, the opening importance since they are the most common of these
of the masonry joint triggered the change in the amplitude types of railway structure.
transfer function. Figure 14 shows that the loaded state 2. Small scale exciter, capable of applying a constant
amplitude surpassed that of the initial state, especially at input force amplitude in whole frequency compo-
low frequency. This indicates that the masonry type retain- nents, was developed as a method to evaluate the
ing wall tends to vibrate easily. structural health of retaining walls.
The model tests demonstrated that the vibration char- 3. Based on the results from the series of model tests
acteristics of the retaining wall were affected by its struc- it has been found out that the vibration characteris-
tural health. In the case of the masonry type retaining tics of a retaining wall are affected by its structural
wall, the importance of the facing rigidity was also high- health.
lighted. 4. The behavior of the retaining wall discussed above
indicates that the spectrum area can be used as
4.3 A proposed procedure to evaluate the structural an index to quantitatively evaluate the structural
health of a retaining wall health of a retaining wall.

The basic concept underlying the proposed method is

schematically illustrated in Fig. 15. As discussed above, 6. Future study
vibration characteristics of the retaining wall were affected
by the structural health of the retaining wall. In the pro- The insufficient of the field data means that there is
posed method, therefore, the area of the amplitude in the no current alternative to comparing spectrum areas of in-

164 QR of RTRI, Vol. 54, No. 3, Aug. 2013

spected retaining walls with those of sound retaining walls
constructed nearby. Site test data shall be collected in the
course of further study in order to be able to determine a
proper threshold value, which may then be applied to dis-
tinguish between deteriorated and sound retaining walls.
6. Future study
The insufficient of the field data means that there is
no current alternative to comparing spectrum areas of in-
spected retaining walls with those of sound retaining walls
constructed nearby. Site test data shall be collected in the
course of further study in order to be able to determine a
proper threshold value, which may then be applied to dis-
tinguish between deteriorated and sound retaining walls.
QR of RTRI, Vol. 54, No. 3, Aug. 2013
View publication stats
References
[1] Railway Technical Research Institute, Maintenance
Standards for Railway structures, 2007.
[2] Oyado, M., Miyashita, M., Ueda, S. and Sakairi, A., An
attempt of rationalization for maintenance of railway
structure using supporting system, Proc. of 5th Inter-
national conference for bridge maintenance, Safety and
Management, pp.3459-3466, 2010.
[3] Nishimura A., Okumura, F. and Tanamura S., Integ-
rity Judgment of Railway Bridges by Percussion Tests
for Structure Response, Quarterly Report of RTRI, Vol.
29, No.4, pp.184-189, 1988.
[4] Nakajima, S., Shinoda, M., Abe, K., Mai, T. and Ehara
T, Application of vibration testing for structural
health evaluation of existing railway retaining walls,
International Workshop on ICT in Geo-Engineering
(ICTGE2012), 2012.
165