- Email: [email protected]
Seismic response of hunchbacked block type gravity quay walls
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Contents lists available at ScienceDirect
Soil Dynamics and Earthquake Engineering journal homepage: www.elsevier.com/locate/soildyn
Seismic response of hunchbacked block type gravity quay walls a
a,⁎
b
Z. Tugce Yuksel , Yalcin Yuksel , K. Onder Cetin , Esin Cevik a b
MARK
a
Yildiz Technical University, Department of Civil Engineering, Davutpasa Campus, Esenler, 34220 Istanbul, Turkey Middle East Technical University, Department of Civil Engineering, Cankaya, 06800 Ankara, Turkey
A R T I C L E I N F O
A B S T R A C T
Keywords: Gravity type quay wall Block type Seismic effects Shaking table Performance based design
Earthquakes near major cities may cause big social and economic impacts. Damages to port facilities may cripple the economy. The past twenty years’ experience has proven the high vulnerability of the port facilities. This fact, along with the economic importance of port structures, indicates the need for better seismic design approaches for berth structures and cargo handling facilities. In the recent decades, there have been many incidences of failure of gravity type quay walls. These failures have stimulated research interest in the development of performance-based design methods. In this paper, two different hunchbacked block type quay walls with different back face shape were studied. A series of 1-g shaking tank tests was performed using a 1/10 scaled block type quay wall with gravel backfill materials on firm non-liquefiable sea bed conditions subjected to different harmonic loads. The shaking tank tests provided insight into the wall displacements and the total dynamic pressures by analyzing pressure components at the contact surface between the saturated gravel backfill soil and the wall. It is concluded that the back-face shape of the walls is an important factor and the larger positive slope of the wall improves the overall seismic stability.
1. Introduction Ports are the main components of maritime transport and they have an important role on commercial transport, hence any damage level is undesirable. Most of the port structures have been located in highly seismic regions and the supporting quay walls are also subjected to earthquake loadings in addition to water wave action and vessels berthing loads. Therefore, the performance of existing port structures should be checked on the basis of earthquake hazard. Seismic risks at ports have not received sufficient attention and only a limited number of studies has been carried out for the assessment of block type quay walls, which are widely preferred in most of the ports. Yuksel et al. [17] documented and discussed the distribution and the extent of damage and serviceability of marine structures after 1999 Kocaeli Earthquake (Mw = 7.52). The effects of earthquakes, including severity of damage, service losses, and environmental impact at petrochemical facilities, were severe and extensive. Sumer et al. [14] presented a state-of-the-art review of seismic-induced liquefaction with special reference to marine structures. The seismic response of a port structure is affected by the interaction of the structure with the surrounding and underlying soil, and water. This effect, widely referred to soil-structure-water interaction (SSWI), is a rather complex phenomenon and involves a number of difficult-to-assess problems. One basic problem is the change in
⁎
Corresponding author. E-mail address: [email protected] (Y. Yuksel).
http://dx.doi.org/10.1016/j.soildyn.2017.08.002 Received 29 June 2017; Accepted 5 August 2017 0267-7261/ © 2017 Elsevier Ltd. All rights reserved.
amplitude and frequency content of seismic waves when they interact with an inclusion in a propagation medium. This kinematic interaction is initiated when incident seismic propagation away from the causative fault and through the geologic media encounter a structural element or foundation element whose inertial and stiffness characteristics differ from those of the surrounding soils. As these incident ground waves hit the structure-foundation, they are both reflected and refracted. The resulting transmitted waves are the source of inertial interaction and generate inertia forces by exciting the overlying structure, which further alter the motions of the foundation and the surrounding soil (ASCE, 1998). Researchers have focused on seismic performance of waterfront structures for longer than a decade and a number of research studies have been carried out both experimentally and numerically. It is very important to learn the lessons from past case studies to better understand the vulnerability of waterfront structures exposed to earthquake. Experimental and/or analytical studies by Miura et al. [9], Fujiwara et al. [3], Mohajeri et al. [10], Mendez et al. [8] and Nakahara et al. [11] were presented to assess the dynamic response of gravity type quay walls. Additionally, Inoue et al. [5], Kim et al. [7], Kim et al. [6], Towhata et al. [16] and Yuksel et al. [18] approached the problem through experimental and/or numerical approaches. There exist also purely numerical studies performed by Alyami et al. [1], Arablouei et al. [2] and Tiznado and Rodriguez-Roa [15]. Most of these studies
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
prepared as shown in Fig. 2. Both models consist of 6 pieces of concrete blocks with a constant block height (h) of 100 mm, width (w) of 250 mm and varying breadths (B) of different aspect ratios designed without shear keys in between blocks. The properties of the concrete blocks are listed in Table 2. Two different overall geometries were obtained by placing six blocks in a different pattern. For the FHW model the breaking point is lower than that of the SHW as illustrated in Fig. 2. The model blocks were founded on an 80 mm thick, tamped gravel seabed floor and numbered from 1 to 6 starting from the bottom block. During the test preparation, utmost attention was given to simulate plane strain conditions. For the purpose of eliminating the effects of tank boundaries on the overall response, models were placed in a plexiglass container, which is fixed to the shaking tank with a gap of 1 mm between the model blocks and the plexi-glass side walls. With the aim of monitoring the performance of the models, series of transducers were installed as shown in Tables 3, 4, and Fig. 2. One accelerometer was placed on the tank, just at the bottom of the models and six accelerometers were fixed onto each block. Six earth pressure and wire type displacement transducers were installed in the back and front sides of the model blocks, respectively. One displacement transducer (DP 7) was installed on top of the sixth block to measure tilting of the wall. Two pore water pressure transducers were embedded in the backfill behind the models at two different locations. The transducers were mounted on each model block, and subsequently the backfill was filled behind the model walls by using pluviation device as shown in Fig. 3. The total mass of the backfill was 810 kg. On the basis of the fact that foundation and backfill soil properties govern the overall response along with the properties of the quay wall, special attention was paid to produce models with desired soil properties. Hence, a series of laboratory tests were performed on foundation and backfill soils to identify their geotechnical engineering properties and the results are summarized in Table 5. The relative density (Dr) of the backfill material was kept constant for each test. To assure a homogeneous relative density, a pluviation device was used. This automated raining crane system is controlled by a digital computer. The device can move both in the vertical and horizontal directions at each pluviation cycle. As shown in Fig. 3, the beam provides the horizontal mobility of the bunker which was connected to the reaction beam by four columns. The backfill material was loaded to the bunker via a conveyor belt as shown in Fig. 4. The relative density of the backfill was selected as 70% for all tests and it is achieved by controlling the discharge cap space, height of the bunker and the speed rate of the horizontal mobility of the bunker. Following the placement of the model blocks and transducers, the backfill material was poured. Then, water was percolated through the bottom of the tank at a very slow rate to avoid boiling and piping induced problems. The maximum free water surface elevation was selected as 68 cm relative to the tank base and backfill soil was fully saturated. For settlement evaluations, the backfill free surface before and after the shaking was measured and recorded using HR Wallingford Touch-Sensitive Two-Dimensional Profiler. Before shaking, the static earth pressures and hydrostatic pressures were monitored to assess the static stress state of the model. Then, four regular harmonic input motions with frequencies ranging from 3 to 7 Hz were applied in the longitudinal direction for 20 s. The characteristics of the harmonics were listed in Table 6. The variations of accelerations, dynamic earth pressures, pore water pressures and displacements were recorded throughout shaking. Once the shaking was ceased, the backfill free surfaces were monitored again for settlement assessment purposes. For the purpose of assessing the performance of the gravity type quay walls, the dimensional parameters are determined as follow:
were focused on understanding the response of caisson type quay walls. Hence, in the literature there exists a gap in understanding the seismic performance of block type quay walls. Sadrekarimi et al. [13] investigated the static and dynamic behavior of hunchbacked gravity walls by considering back-face shape of wall. Sadrekarimi [12] also studied the seismic performance of gravity type broken back quay walls through 1 g shaking table model experiments and proposed a simplified sliding block analysis model for estimating lateral displacements, which is calibrated with the experimental results. However, the blocks had shear keys at the top and bottom surfaces to prevent relative sliding. A contemporary design philosophy for port structures in seismically active regions is expected to suggest assessment methodologies for the estimation of seismically induced foundation, backfill and wall deformations along with the stresses acting on them. Unfortunately, conventional (force-balance) methods are not well suited to fulfill these expectations. While performance-based design procedures attempt to assess the deformation and stress demand and capacity of the systems however appropriate earthquake performance levels needed to be defined along with acceptable block type quay wall damages. Despite their wide use as a quay wall, in the literature there exists a gap on acceptable performance criteria. This study attempts to assess the static and dynamic performance of the block type quay walls in the form of lateral displacement and tilting as well as settlement of the backfill and is hoped to contribute to close this gap. In this study, reduced scale models of two different hunchbacked quay walls with different back face shapes were prepared in 1-g shaking tank. The scale ratio of model to prototype was selected as 1/10. Tests were performed on firm bottom conditions and a dense backfill material was used for the purpose of fully concentrating on the response of quay wall and eliminating the effects of soil liquefaction on the overall response. 2. Experimental study Due to the fact that gravity type water-front structures are usually long, one-dimensional dynamic assessments are widely used to understand the overall response of these structures. 1-g model tests with a 1/ 10 scale ratio were performed with the similitude of various parameters as recommended by [4] and given in Table 1, for a soil-structure-fluid system. The experimental study was conducted in a shaking tank at Hydraulic and Coastal Engineering Laboratory of Yildiz Technical University. As shown in Fig. 1, the shaking tank is 4.5 m long, 1 m wide and 1 m high with a glass side to monitor the overall response. The tank was divided by a steel plate in the longitudinal direction and 0.38 m wide models were centered in the tank as shown in Fig. 1. The shaking tank was steered by PLC (Programmable Logic Control). The amplitude and frequency ranges of the tank is 1–5 mm and 1–9 Hz, respectively. Series of calibration tests were performed to develop the optimum amplitude and frequency pair to attain reliable and robust response. For the purpose of investigating the performance of hunchbacked block type gravity walls, two different models; i) first type hunchbacked wall (FHW), and ii) second type hunchbacked wall (SHW), were Table 1 Similitude for the 1-g shaking table model [4].
Length Time Acceleration Displacement Water Pressure Density Stress
Prototype/model
Scale factor
λ λ0.75 1 λ1.5 λ 1 λ
10 5.62 1 31.62 10 1 10
F(ρs , d50 , H, B, ρc , ρw , a, g, D, ZB, θ) = 0
(1)
where; D is the maximum residual displacement of the block, B is the block breadth, H is the wall height, ρc , ρs and ρw are the densities of the 226
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 1. Shaking tank (all dimensions in mm).
Fig. 2. Model profiles.
block, backfill soil and water, respectively, d50 is the medium grain-size of the backfill soil, a is the acceleration recorded at each block, g is the gravitational acceleration, ZB is the settlement of the backfill soil and θ is the tilt angle respect to vertical plane. The dimensionless parameters, useful for assessing the performance of the gravity quay wall were determined by using dimensional analysis, which are listed as follow:
Table 2 Properties of the concrete blocks. Dimensions (h × B × w) (cm × cm × cm)
Mass (kg)
Density (kg/dm3)
Number of pieces
10 10 10 10
10.80 13.30 16.20 18.45
2.16 2.13 2.16 2.11
1 2 2 1
× × × ×
20 25 30 35
× × × ×
25 25 25 25
a Z D B ρ ρ d f ⎛⎜ , , c , w , 50 , , B , θ⎞⎟ = 0 H H ρs ρs H g H ⎠ ⎝
where D/H, B/H are the normalized lateral displacement and normalized the gravity wall dimension terms, whereas ρc /ρs and ρw /ρs are the normalized density terms, d50 /H is the normalized grain-size, a/g = a (g) is the seismic acceleration in gravitational acceleration units, ZB/H is the term representing the relative settlement of the backfill, and θ is the tilt angle. The dimensionless function can be further simplified as given in Eq. (3), since ρc /ρs , ρw /ρs and d50 /H terms are constant for a given model.
Table 3 Properties of the monitoring devices. Accelerometers (AC)
Pore Pressure Transducers (PP) Total Soil Pressure Transducers (SP) Displacement Transducers (DP)
(2)
IMI Sensors type 626B13 ICP, 1000 mV/g sensitivity, ± 5 g measuring range, 0.2–6000 Hz frequency range and ± %1 accuracy KPSI sensors, ± 0.34–20.34 mH2O Sokki Kenkyujo Systems type KDG-200KPA
D B D a Z f ⎛⎜ , , , , B , θ⎞⎟ = 0 ⎝H H B g H ⎠
UniMeasure Systems type HX-PA, 0.1% sensitivity
(3)
In the study, the backfill material was selected as gravel, which is judged to be non-liquefiable. Pore pressure transducers installed within 227
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Table 4 Locations of monitoring devices. Accelerometers (AC)
Soil pressure transducers (SP)
Displacement transducers (DP)
AC1 AC2 AC3 AC4 AC5 AC6 AC7
– SP1 SP2 SP3 SP4 SP5 SP6
– DP1 DP2 DP3 DP4 DP5 DP6 DP7a
Pore water pressure transducers (PP)
PP1 PP2 a
Location
Tank Base Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 6 (top of the walls) 20 cm from bottom in the backfill 40 cm from bottom in the backfill
for vertical displacement measurement.
block (water side) and they were labeled as AC2 through AC7, indices increasing from bottom to top. As discussed earlier, the duration of the harmonics was selected as 20 s and a sample base acceleration-time history was given in Fig. 5. To eliminate the effects of small deviations in the amplitude, the averages of the harmonic base acceleration records (aavg,b) were estimated for each test and given in Table 6. The average acceleration of each block (aavg,x)was normalized with the average base acceleration (aavg,b) to estimate the relative amplification/ deamplification ratios (Am,avg) of each block. The subscript “x” in aavg,x expresses the location of the accelerometer. The scatters of Am,avg with aavg,b for each block of both models were shown in Fig. 6. As revealed by the figure, the average amplification ratio (and normalized inertial forces) increased with increasing average base acceleration for both hunchbacked walls. Additionally, the maximum of average amplification ratios (i.e.: normalized inertia forces) was observed at the upper most blocks, valid for all tests. The amplifications recorded for the SHW were observed to be larger than those of the FHW. The average amplification ratios (and normalized inertial forces) recorded for the top block of SHW were observed to be approximately 75% larger than those of FHW. Hence, from inertial forces point of view, the FHW as compared to SHW performed significantly better and were subjected to less normalized inertial forces.
Fig. 3. Schematic view of the pluviation device.
Table 5 General properties of the backfill.
3.2. Earth pressure
Parameter Void ratio, (e) Porosity, (n) Dry unit weight, (γk) Saturated unit weight, (γd) Specific gravity, (Gs) Average grain diameter, (D50) Effective grain diameter, (D10) The uniformity coefficient, (Cu) The gradation coefficient, (Cc)
As discussed earlier a total of 6 earth pressure transducers were installed immediately behind each block, the locations of which were shown in Figs. 1 and 2. The recorded pressure-time histories decomposed into two components and were named as non-oscillating and oscillating. The non-oscillating component of the total lateral earth pressure was obtained by time weighted average, and the oscillating component was determined by deducting the non-oscillating component from the total earth pressure. Fig. 7 schematically illustrated how these components were estimated for Block # 1 and #6. As revealed by this figure, the total earth pressure acting on Block # 1 was sharply reduced in the first couple of seconds of shaking, then reached to a stability. Contrary this response of Block # 1, the total lateral earth pressure acting on Block # 6 gradually increased until the 5th second of the loading, again followed by a stabilized response. Similar behavioral trends were observed in both hunchbacked walls. Oscillating earth pressures recorded at different elevations enabled to identify if and when phase differences in responses were observed between block and backfill soil. This was illustrated in Fig. 8, where the oscillating earth pressure and the acceleration of Block # 3 were comparatively shown with zoom in view for the time between 9th and 12th seconds. The oscillating components of the earth pressure and acceleration of the block were observed in-phase. This observation is valid for all blocks of both of the hunchbacked walls. The variation of lateral earth pressures with depth was presented in Fig. 9 for both of the hunchbacked wall configurations. The maximum
0.631 0.386 16.413 (kN/m3) 20.230 (kN/m3) 2.730 0.825 (cm) 0.500 (cm) 1.8 1.08
the gravel backfill also confirmed this assumption. Thus, dynamic water forces acting on the wall were assumed to be purely due to shaking of the soil and water mass and not due to excess pore water pressure generation in the backfill. Lateral earth pressures, accelerations of the rigid model blocks (i.e.: inertia forces), dynamic water pressure and frictional forces between foundation and model blocks were monitored throughout shaking for different back shape of gravity quay walls. 3. Results and discussion 3.1. Acceleration-time histories The base acceleration was labeled as AC1 as shown in Fig. 2. Additional accelerometers were installed at the center of front face of each 228
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 4. Shaking tank, pluviation device and the conveyor belt.
Table 6 Characteristics of shaking. Test no
1 2 3 4
Frequency (Hz)
3.0 7.0 4.0 4.0
Base acceleration (g) FHW
SHW
0.063 0.080 0.219 0.346
0.062 0.0827 0.216 0.349
Fig. 5. The base acceleration-time history for Test No: 3. Fig. 7. Total earth pressure and its two components for the bottom and top blocks of the FHW, Test No: 3 (a) Block 1, (b) Block 6.
Fig. 6. Block amplifications for the harmonic motions of (a) FHW and (b) SHW.
accelerations were observed to be linearly amplified when propagating from the bottom to top for maximum base acceleration amplitudes less than 0.2 g. For base acceleration exceeding 0.2 g, nonlinearly amplifying acceleration levels were observed and additionally the back-face
Fig. 8. Comparison between the oscillating earth pressures and the acceleration of the block for the FWH, Test No: 2, Block 3.
229
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 9. Distribution of the total earth pressure with depth for aavg,b < 0.2 g (Test No: 2).
Fig. 11. Pore water pressure ratios for FHW, (Test No: 3).
3.4. Displacement The displacement transducers were installed on each block of both model walls to determine sliding and tilting responses of the walls as shown earlier in Fig. 2. The horizontal displacements of each block (DP1 through DP6) were monitored at the center of the blocks whereas the vertical displacements (DP7) were measured at the top of the walls. Despite firm bottom conditions, small tilt angles were measured towards the water side under strong shaking without sinking into the seabed (Fig. 12). A comparative evaluation indicated that the tilt angle of SHW was about 1.5 times larger than that of FHW at the largest base accelerations corresponding to test 4. Illustrative horizontal displacement - time histories of each block of FHW and SHW models are shown in Fig. 13a and b, respectively. As revealed by the figures, the horizontal displacements increase with time, during the test duration and also block displacements increased with time and elevation. The normalized horizontal displacements (Dx/ z) recorded at the 10th second of the shaking versus base acceleration pairs were given in Fig. 14 for each wall type. It should be noted that z is the elevation difference of the displacement transducers relative to the backfill surface. It was clearly observed that the relative displacements increased with increasing base acceleration. Moreover, Fig. 15 showed that, the relative horizontal displacement was larger, the extent of which reached to a maximum of 1.8 times for SHW as compared to the ones of FHW model. The differences in the responses were attributed to a down-drag forces and the weight of the soil block resting on the widest block. In simpler terms, during model preparation, the backfill soil settled under its own weight, and mobilized a down-drag force, which mobilized some friction between the wall and the backfill soil. This frictional force and the weight of the soil resting on the widest positive back-slope of FHW stabilized the wall and increased the resistance against overturning and sliding. Since, the frictional surfaces of each block of the hunchbacked walls are different, the normalized horizontal displacements with block breadths were also considered. The results were compared for both hunchbacked walls as shown in Fig. 16. The normalized displacements (Dx/B) were relatively larger for SHW. The normalized displacements of the widest block (i.e.: the broken point of the walls) were found to be relatively small when compared with the other block displacements for both models. The differences in normalized displacements between two model types reached to a maximum of 33% at the top block for Test No. 3.
Fig. 10. Distribution of the total earth pressure with depth for aavg,b > 0.2 g, (a) FHW and (b) SHW (Test No: 4).
shape of the wall especially for both hunchbacked walls became an important factor of the dynamic interaction between the backfill soil and the wall. As shown in Fig. 10, for base acceleration amplitudes exceeding 0.2 g, lateral earth pressures were observed to increase linearly till the widest block, then a sudden decrease was observed at depths immediately below the widest block. This sudden decrease in lateral earth pressure was attributed to shadowing effects of the widest block.
3.3. Pore water pressure To better present the pore pressure responses of the backfill soil, a dimensionless parameter: excess pore water pressure ratio, ru was defined as given in Eq. (4).
ru =
Δu ′ σv,0
(4)
where Δu is the excess pore pressure, and σ′v,0 is the initial vertical effective stress. When Δu becomes equal to the initial vertical effective stress, then instantaneously vertical effective stress becomes zero during cyclic loading leading to ru value of 100%. Since shear strength in cohesionless soils is proportional with normal effective stress, an ru value of 100% suggest very low shear strength values. This phenomenon is widely referred to as soil liquefaction. In this study, the pore pressure transducers, PP1 and PP2 were installed 20 cm apart from each other for both wall types as shown earlier in Fig. 2. Due to high hydraulic conductivity of the coarse backfill soil, the excess pore water pressures developed during shaking were observed to be negligibly small as shown in Fig. 11.
230
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 12. Schematization of the tilting behavior of the (a) first and second-type hunchbacked walls and (b) their tilting responses, respectively.
Fig. 13. Horizontal displacement-time histories of both hunchbacked walls (Test No: 3).
Fig. 14. Relative horizontal displacements (t = 10 s) for (a) FHW and (b) SHW.
3.5. Settlement
the total depth of the backfill soil (HB), and these values vs. base accelerations were shown in Fig. 18. As revealed by the figure, the normalized settlements increased with increasing base acceleration amplitudes addressing the nonlinear response of backfill soil. Comparing the results for both hunchbacked walls, especially for larger base accelerations (aavg,b > 0.2 g), the normalized settlements were relatively smaller for FHW. This response is rather expected since smaller
The surface settlement profiles of the backfill were developed by monitoring the surface elevations before and after shaking. As shown in Fig. 17, it was observed that the maximum settlement (ZB) was observed to be not located immediately behind the wall but ¼ H away from the walls and the settlements for SHW were observed to be larger than that of FHW. The maximum settlement of each test was divided by 231
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
during and after shaking wall displacements, total earth and pore pressures were monitored. Additionally, after the end of shaking, the soil profiles behind the model walls were also monitored. Followings are the specific conclusions of these studies: 1. From inertial forces point of view, the FHW as compared to SHW showed a significantly better performance level and are subjected to less normalized inertial forces. 2. Consistently, the average amplification ratios recorded at the top block of SHW were observed to be approximately 75% larger than those of FHW. 3. The oscillating component of the earth pressure and accelerations of the blocks were observed to be in-phase, which is valid for all blocks of both of the hunchbacked walls. 4. The variation of lateral earth pressures with depth were observed to be linear for maximum base acceleration amplitudes less than 0.2 g. 5. While for base acceleration exceeding 0.2 g, nonlinearly amplifying acceleration levels were observed and the back-face shape of the wall (i.e.: the wall type) became an important factor of the dynamic interaction between the backfill soil and the wall. 6. The observation of horizontal displacement, tilting and settlement results revealed that the first type hunchbacked wall exhibited a more stable performance as compared to the performance of the second type hunchbacked wall, especially at the base acceleration levels exceeding 0.2 g. 7. Consistently, the maximum tilt angle and the normalized horizontal displacements of SHW were observed to be about 1.5 and 1.8 times larger than that of FHW, respectively. 8. The settlements were also observed to be larger for SHW.
Fig. 15. Comparison of the relative horizontal displacements (Dx/Z) for both hunchbacked walls (Test No: 3).
Fig. 16. Comparison of the normalized horizontal displacements (Dx/B) for both hunchbacked walls (Test No: 3).
As the concluding remark, the seismic performance of these hunchbacked walls is controlled by their different back-face geometries, and larger positive hunched walls (i.e.: FHW) perform significantly better under seismic excitation leading to more economical and safer design.
Fig. 17. Results of the profile measurements for both hunchbacked walls (Test No: 3).
Acknowledgement We would like to thank to Scientific and Technological Research Council of Turkey (TUBITAK) Grant No: 113M426. References [1] Alyami M, Rouainia M, Wilkinson SM. Numerical analysis of deformation behavior of quay walls under earthquake loading. J Soil Dyn Earthq Eng 2009;29:525–36. [2] Arablouei A, Gharabaghi ARM, Ghalandarzadeh A, Abedi K, Ishibashi I. Effects of seawater-structure-soil interaction on seismic performance of caisson type quay wall. Comput Struct 2011;89:2439–59. [3] Fujiwara T, Horikoshi K, Higuchi Y, Sueoka T. Estimation of dynamic displacement of gravity type quay walls based on centrifuge modelling, In: Proceedings of the 12th world conference on earthquake engineering, Japan. [4] Iai S. Similitude for shaking table tests on soil-structure-fluid in 1g gravitational field. Soils Found 1989;29(1):105–18. [5] Inoue E, Miura K, Otsuka N, Yoshida N, Sasajima T. Numerical analysis of the earth pressure during earthquake on the gravity type quay wall, In: Proceedings of the thirtheenth (2003) international offshore and polar engineering conference, 2003, 25–30 May 2003, Hawai. [6] Kim SR, Jang IS, Chung CK, Kim MM. Evaluation of seismic displacements of quay walls. J Soil Dyn Earthq Eng 2005;25:451–9. [7] Kim SR, Kwon OS, Kim MM. Evaluation of force components acting on gravity type quay walls during earthquakes. J Soil Dyn Earthq Eng 2004;24:853–66. [8] Mendez BC, Botero E, Romo MP. A new friction law for sliding rigid blocks under cyclic loading. Soil Dyn Earthq Eng 2008;29:874–82. [9] Miura K, Kohama E, Inoue K, Otsuka N, Sasajima T, Hayashi T, Yoshida N. Behaviour of gravity type quay wall during earthquake regarding dynamic interaction between qaisson and backfill during liquefaction, In: Proceedings of the 12th world conference on earthquake engineering, Japan. [10] Mohajeri M, Ichii K, Tamura T. Modification of the sliding block concept for caisson walls. J Waterw, Port, Coast Ocean Eng, ASCE 2002;134:134–42. [11] Nakahara T, Kohama E, Sugano T. Model shake table test on the seismic performance of gravity type quay wall with different foundation ground properties, In: Proceedings of the 13th world conference on earthquake engineering (13WCEE),
Fig. 18. Normalized settlements with respect to base accelerations for both hunchbacked walls.
horizontal displacements of the wall lead to reduced settlements behind the wall. Also, the larger positive slope of the wall was believed to stabilize the wall subjected to shaking. On the basis of this, it can be concluded that first type hunchbacked wall has a zone immediately behind the wall, where it is safer to locate industrial and coastal structures. This observation is of value to reduce the damages to cranes and pavements, if they are located in this less settled zone behind the quay walls. 4. Summary and conclusions Within the confines of this manuscript, seismic responses of two different hunchbacked block type quay walls with different back face shape were presented. A series of 1-g shaking tank tests was performed using a 1/10 scaled block type quay wall with gravel backfill materials on firm non-liquefiable sea bed conditions. Sinusoidal harmonic shakings with frequencies ranging from 3 Hz to 7 Hz, and average base accelerations in the range of 0.062 g to 0.349 g were applied. Before, 232
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
[15] Tiznado JC, Rodriguez-Roa F. Seismic lateral movement prediction for gravity retaining walls on granular soils. Soil Dyn Earthq Eng 2011;31:391–400. [16] Towhata I, Alam MJ, Honda T, Tamate S. Model tests on behaviour of gravity type quay walls subjected to strong shaking. Bull NZ Soc Earthq Eng 2009;42(1):47–56. [17] Yuksel Y, Alpar B, Yalciner AC, Cevik E, Ozguven O, Celikoglu Y. Effects of Eastern Marmara earthquake on the marine structures and coastal areas. J Water Marit Eng, ICE 2002;156:147–63. [18] Yuksel Y, Yuksel ZT, Cevik E, Orhan K, Berilgen M. Evaluation of the seismic performance of a caisson and L-type quay wall. Soil Dyn Earthq Eng 2017;92:537–50.
2004, 1–6 August 2004, Canada. [12] Sadrekarimi A. Seismic displacement of broken-back gravity quay walls. J Waterw, Port, Coast Ocean Eng, ASCE 2011;137/2:75–84. [13] Sadrekarimi A, Ghalandarzadeh A, Sadrekarimi J. Static and dynamic behaviour of hunchbacked gravity quay walls. Soil Dyn Earthq Eng 2008;28:99–117. [14] Sumer M, Ansal A, Cetin KO, Damgaard J, Gunbak AR, Hansen NEO, Sawicki A, Synolakis CE, Yalciner AC, Yuksel Y, Zen K. Earthquake-induced liquefaction around marine structures. J Waterw, Port, Coast Ocean Eng, ASCE 2007;133(1):55–82.
233
Contents lists available at ScienceDirect
Soil Dynamics and Earthquake Engineering journal homepage: www.elsevier.com/locate/soildyn
Seismic response of hunchbacked block type gravity quay walls a
a,⁎
b
Z. Tugce Yuksel , Yalcin Yuksel , K. Onder Cetin , Esin Cevik a b
MARK
a
Yildiz Technical University, Department of Civil Engineering, Davutpasa Campus, Esenler, 34220 Istanbul, Turkey Middle East Technical University, Department of Civil Engineering, Cankaya, 06800 Ankara, Turkey
A R T I C L E I N F O
A B S T R A C T
Keywords: Gravity type quay wall Block type Seismic effects Shaking table Performance based design
Earthquakes near major cities may cause big social and economic impacts. Damages to port facilities may cripple the economy. The past twenty years’ experience has proven the high vulnerability of the port facilities. This fact, along with the economic importance of port structures, indicates the need for better seismic design approaches for berth structures and cargo handling facilities. In the recent decades, there have been many incidences of failure of gravity type quay walls. These failures have stimulated research interest in the development of performance-based design methods. In this paper, two different hunchbacked block type quay walls with different back face shape were studied. A series of 1-g shaking tank tests was performed using a 1/10 scaled block type quay wall with gravel backfill materials on firm non-liquefiable sea bed conditions subjected to different harmonic loads. The shaking tank tests provided insight into the wall displacements and the total dynamic pressures by analyzing pressure components at the contact surface between the saturated gravel backfill soil and the wall. It is concluded that the back-face shape of the walls is an important factor and the larger positive slope of the wall improves the overall seismic stability.
1. Introduction Ports are the main components of maritime transport and they have an important role on commercial transport, hence any damage level is undesirable. Most of the port structures have been located in highly seismic regions and the supporting quay walls are also subjected to earthquake loadings in addition to water wave action and vessels berthing loads. Therefore, the performance of existing port structures should be checked on the basis of earthquake hazard. Seismic risks at ports have not received sufficient attention and only a limited number of studies has been carried out for the assessment of block type quay walls, which are widely preferred in most of the ports. Yuksel et al. [17] documented and discussed the distribution and the extent of damage and serviceability of marine structures after 1999 Kocaeli Earthquake (Mw = 7.52). The effects of earthquakes, including severity of damage, service losses, and environmental impact at petrochemical facilities, were severe and extensive. Sumer et al. [14] presented a state-of-the-art review of seismic-induced liquefaction with special reference to marine structures. The seismic response of a port structure is affected by the interaction of the structure with the surrounding and underlying soil, and water. This effect, widely referred to soil-structure-water interaction (SSWI), is a rather complex phenomenon and involves a number of difficult-to-assess problems. One basic problem is the change in
⁎
Corresponding author. E-mail address: [email protected] (Y. Yuksel).
http://dx.doi.org/10.1016/j.soildyn.2017.08.002 Received 29 June 2017; Accepted 5 August 2017 0267-7261/ © 2017 Elsevier Ltd. All rights reserved.
amplitude and frequency content of seismic waves when they interact with an inclusion in a propagation medium. This kinematic interaction is initiated when incident seismic propagation away from the causative fault and through the geologic media encounter a structural element or foundation element whose inertial and stiffness characteristics differ from those of the surrounding soils. As these incident ground waves hit the structure-foundation, they are both reflected and refracted. The resulting transmitted waves are the source of inertial interaction and generate inertia forces by exciting the overlying structure, which further alter the motions of the foundation and the surrounding soil (ASCE, 1998). Researchers have focused on seismic performance of waterfront structures for longer than a decade and a number of research studies have been carried out both experimentally and numerically. It is very important to learn the lessons from past case studies to better understand the vulnerability of waterfront structures exposed to earthquake. Experimental and/or analytical studies by Miura et al. [9], Fujiwara et al. [3], Mohajeri et al. [10], Mendez et al. [8] and Nakahara et al. [11] were presented to assess the dynamic response of gravity type quay walls. Additionally, Inoue et al. [5], Kim et al. [7], Kim et al. [6], Towhata et al. [16] and Yuksel et al. [18] approached the problem through experimental and/or numerical approaches. There exist also purely numerical studies performed by Alyami et al. [1], Arablouei et al. [2] and Tiznado and Rodriguez-Roa [15]. Most of these studies
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
prepared as shown in Fig. 2. Both models consist of 6 pieces of concrete blocks with a constant block height (h) of 100 mm, width (w) of 250 mm and varying breadths (B) of different aspect ratios designed without shear keys in between blocks. The properties of the concrete blocks are listed in Table 2. Two different overall geometries were obtained by placing six blocks in a different pattern. For the FHW model the breaking point is lower than that of the SHW as illustrated in Fig. 2. The model blocks were founded on an 80 mm thick, tamped gravel seabed floor and numbered from 1 to 6 starting from the bottom block. During the test preparation, utmost attention was given to simulate plane strain conditions. For the purpose of eliminating the effects of tank boundaries on the overall response, models were placed in a plexiglass container, which is fixed to the shaking tank with a gap of 1 mm between the model blocks and the plexi-glass side walls. With the aim of monitoring the performance of the models, series of transducers were installed as shown in Tables 3, 4, and Fig. 2. One accelerometer was placed on the tank, just at the bottom of the models and six accelerometers were fixed onto each block. Six earth pressure and wire type displacement transducers were installed in the back and front sides of the model blocks, respectively. One displacement transducer (DP 7) was installed on top of the sixth block to measure tilting of the wall. Two pore water pressure transducers were embedded in the backfill behind the models at two different locations. The transducers were mounted on each model block, and subsequently the backfill was filled behind the model walls by using pluviation device as shown in Fig. 3. The total mass of the backfill was 810 kg. On the basis of the fact that foundation and backfill soil properties govern the overall response along with the properties of the quay wall, special attention was paid to produce models with desired soil properties. Hence, a series of laboratory tests were performed on foundation and backfill soils to identify their geotechnical engineering properties and the results are summarized in Table 5. The relative density (Dr) of the backfill material was kept constant for each test. To assure a homogeneous relative density, a pluviation device was used. This automated raining crane system is controlled by a digital computer. The device can move both in the vertical and horizontal directions at each pluviation cycle. As shown in Fig. 3, the beam provides the horizontal mobility of the bunker which was connected to the reaction beam by four columns. The backfill material was loaded to the bunker via a conveyor belt as shown in Fig. 4. The relative density of the backfill was selected as 70% for all tests and it is achieved by controlling the discharge cap space, height of the bunker and the speed rate of the horizontal mobility of the bunker. Following the placement of the model blocks and transducers, the backfill material was poured. Then, water was percolated through the bottom of the tank at a very slow rate to avoid boiling and piping induced problems. The maximum free water surface elevation was selected as 68 cm relative to the tank base and backfill soil was fully saturated. For settlement evaluations, the backfill free surface before and after the shaking was measured and recorded using HR Wallingford Touch-Sensitive Two-Dimensional Profiler. Before shaking, the static earth pressures and hydrostatic pressures were monitored to assess the static stress state of the model. Then, four regular harmonic input motions with frequencies ranging from 3 to 7 Hz were applied in the longitudinal direction for 20 s. The characteristics of the harmonics were listed in Table 6. The variations of accelerations, dynamic earth pressures, pore water pressures and displacements were recorded throughout shaking. Once the shaking was ceased, the backfill free surfaces were monitored again for settlement assessment purposes. For the purpose of assessing the performance of the gravity type quay walls, the dimensional parameters are determined as follow:
were focused on understanding the response of caisson type quay walls. Hence, in the literature there exists a gap in understanding the seismic performance of block type quay walls. Sadrekarimi et al. [13] investigated the static and dynamic behavior of hunchbacked gravity walls by considering back-face shape of wall. Sadrekarimi [12] also studied the seismic performance of gravity type broken back quay walls through 1 g shaking table model experiments and proposed a simplified sliding block analysis model for estimating lateral displacements, which is calibrated with the experimental results. However, the blocks had shear keys at the top and bottom surfaces to prevent relative sliding. A contemporary design philosophy for port structures in seismically active regions is expected to suggest assessment methodologies for the estimation of seismically induced foundation, backfill and wall deformations along with the stresses acting on them. Unfortunately, conventional (force-balance) methods are not well suited to fulfill these expectations. While performance-based design procedures attempt to assess the deformation and stress demand and capacity of the systems however appropriate earthquake performance levels needed to be defined along with acceptable block type quay wall damages. Despite their wide use as a quay wall, in the literature there exists a gap on acceptable performance criteria. This study attempts to assess the static and dynamic performance of the block type quay walls in the form of lateral displacement and tilting as well as settlement of the backfill and is hoped to contribute to close this gap. In this study, reduced scale models of two different hunchbacked quay walls with different back face shapes were prepared in 1-g shaking tank. The scale ratio of model to prototype was selected as 1/10. Tests were performed on firm bottom conditions and a dense backfill material was used for the purpose of fully concentrating on the response of quay wall and eliminating the effects of soil liquefaction on the overall response. 2. Experimental study Due to the fact that gravity type water-front structures are usually long, one-dimensional dynamic assessments are widely used to understand the overall response of these structures. 1-g model tests with a 1/ 10 scale ratio were performed with the similitude of various parameters as recommended by [4] and given in Table 1, for a soil-structure-fluid system. The experimental study was conducted in a shaking tank at Hydraulic and Coastal Engineering Laboratory of Yildiz Technical University. As shown in Fig. 1, the shaking tank is 4.5 m long, 1 m wide and 1 m high with a glass side to monitor the overall response. The tank was divided by a steel plate in the longitudinal direction and 0.38 m wide models were centered in the tank as shown in Fig. 1. The shaking tank was steered by PLC (Programmable Logic Control). The amplitude and frequency ranges of the tank is 1–5 mm and 1–9 Hz, respectively. Series of calibration tests were performed to develop the optimum amplitude and frequency pair to attain reliable and robust response. For the purpose of investigating the performance of hunchbacked block type gravity walls, two different models; i) first type hunchbacked wall (FHW), and ii) second type hunchbacked wall (SHW), were Table 1 Similitude for the 1-g shaking table model [4].
Length Time Acceleration Displacement Water Pressure Density Stress
Prototype/model
Scale factor
λ λ0.75 1 λ1.5 λ 1 λ
10 5.62 1 31.62 10 1 10
F(ρs , d50 , H, B, ρc , ρw , a, g, D, ZB, θ) = 0
(1)
where; D is the maximum residual displacement of the block, B is the block breadth, H is the wall height, ρc , ρs and ρw are the densities of the 226
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 1. Shaking tank (all dimensions in mm).
Fig. 2. Model profiles.
block, backfill soil and water, respectively, d50 is the medium grain-size of the backfill soil, a is the acceleration recorded at each block, g is the gravitational acceleration, ZB is the settlement of the backfill soil and θ is the tilt angle respect to vertical plane. The dimensionless parameters, useful for assessing the performance of the gravity quay wall were determined by using dimensional analysis, which are listed as follow:
Table 2 Properties of the concrete blocks. Dimensions (h × B × w) (cm × cm × cm)
Mass (kg)
Density (kg/dm3)
Number of pieces
10 10 10 10
10.80 13.30 16.20 18.45
2.16 2.13 2.16 2.11
1 2 2 1
× × × ×
20 25 30 35
× × × ×
25 25 25 25
a Z D B ρ ρ d f ⎛⎜ , , c , w , 50 , , B , θ⎞⎟ = 0 H H ρs ρs H g H ⎠ ⎝
where D/H, B/H are the normalized lateral displacement and normalized the gravity wall dimension terms, whereas ρc /ρs and ρw /ρs are the normalized density terms, d50 /H is the normalized grain-size, a/g = a (g) is the seismic acceleration in gravitational acceleration units, ZB/H is the term representing the relative settlement of the backfill, and θ is the tilt angle. The dimensionless function can be further simplified as given in Eq. (3), since ρc /ρs , ρw /ρs and d50 /H terms are constant for a given model.
Table 3 Properties of the monitoring devices. Accelerometers (AC)
Pore Pressure Transducers (PP) Total Soil Pressure Transducers (SP) Displacement Transducers (DP)
(2)
IMI Sensors type 626B13 ICP, 1000 mV/g sensitivity, ± 5 g measuring range, 0.2–6000 Hz frequency range and ± %1 accuracy KPSI sensors, ± 0.34–20.34 mH2O Sokki Kenkyujo Systems type KDG-200KPA
D B D a Z f ⎛⎜ , , , , B , θ⎞⎟ = 0 ⎝H H B g H ⎠
UniMeasure Systems type HX-PA, 0.1% sensitivity
(3)
In the study, the backfill material was selected as gravel, which is judged to be non-liquefiable. Pore pressure transducers installed within 227
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Table 4 Locations of monitoring devices. Accelerometers (AC)
Soil pressure transducers (SP)
Displacement transducers (DP)
AC1 AC2 AC3 AC4 AC5 AC6 AC7
– SP1 SP2 SP3 SP4 SP5 SP6
– DP1 DP2 DP3 DP4 DP5 DP6 DP7a
Pore water pressure transducers (PP)
PP1 PP2 a
Location
Tank Base Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 6 (top of the walls) 20 cm from bottom in the backfill 40 cm from bottom in the backfill
for vertical displacement measurement.
block (water side) and they were labeled as AC2 through AC7, indices increasing from bottom to top. As discussed earlier, the duration of the harmonics was selected as 20 s and a sample base acceleration-time history was given in Fig. 5. To eliminate the effects of small deviations in the amplitude, the averages of the harmonic base acceleration records (aavg,b) were estimated for each test and given in Table 6. The average acceleration of each block (aavg,x)was normalized with the average base acceleration (aavg,b) to estimate the relative amplification/ deamplification ratios (Am,avg) of each block. The subscript “x” in aavg,x expresses the location of the accelerometer. The scatters of Am,avg with aavg,b for each block of both models were shown in Fig. 6. As revealed by the figure, the average amplification ratio (and normalized inertial forces) increased with increasing average base acceleration for both hunchbacked walls. Additionally, the maximum of average amplification ratios (i.e.: normalized inertia forces) was observed at the upper most blocks, valid for all tests. The amplifications recorded for the SHW were observed to be larger than those of the FHW. The average amplification ratios (and normalized inertial forces) recorded for the top block of SHW were observed to be approximately 75% larger than those of FHW. Hence, from inertial forces point of view, the FHW as compared to SHW performed significantly better and were subjected to less normalized inertial forces.
Fig. 3. Schematic view of the pluviation device.
Table 5 General properties of the backfill.
3.2. Earth pressure
Parameter Void ratio, (e) Porosity, (n) Dry unit weight, (γk) Saturated unit weight, (γd) Specific gravity, (Gs) Average grain diameter, (D50) Effective grain diameter, (D10) The uniformity coefficient, (Cu) The gradation coefficient, (Cc)
As discussed earlier a total of 6 earth pressure transducers were installed immediately behind each block, the locations of which were shown in Figs. 1 and 2. The recorded pressure-time histories decomposed into two components and were named as non-oscillating and oscillating. The non-oscillating component of the total lateral earth pressure was obtained by time weighted average, and the oscillating component was determined by deducting the non-oscillating component from the total earth pressure. Fig. 7 schematically illustrated how these components were estimated for Block # 1 and #6. As revealed by this figure, the total earth pressure acting on Block # 1 was sharply reduced in the first couple of seconds of shaking, then reached to a stability. Contrary this response of Block # 1, the total lateral earth pressure acting on Block # 6 gradually increased until the 5th second of the loading, again followed by a stabilized response. Similar behavioral trends were observed in both hunchbacked walls. Oscillating earth pressures recorded at different elevations enabled to identify if and when phase differences in responses were observed between block and backfill soil. This was illustrated in Fig. 8, where the oscillating earth pressure and the acceleration of Block # 3 were comparatively shown with zoom in view for the time between 9th and 12th seconds. The oscillating components of the earth pressure and acceleration of the block were observed in-phase. This observation is valid for all blocks of both of the hunchbacked walls. The variation of lateral earth pressures with depth was presented in Fig. 9 for both of the hunchbacked wall configurations. The maximum
0.631 0.386 16.413 (kN/m3) 20.230 (kN/m3) 2.730 0.825 (cm) 0.500 (cm) 1.8 1.08
the gravel backfill also confirmed this assumption. Thus, dynamic water forces acting on the wall were assumed to be purely due to shaking of the soil and water mass and not due to excess pore water pressure generation in the backfill. Lateral earth pressures, accelerations of the rigid model blocks (i.e.: inertia forces), dynamic water pressure and frictional forces between foundation and model blocks were monitored throughout shaking for different back shape of gravity quay walls. 3. Results and discussion 3.1. Acceleration-time histories The base acceleration was labeled as AC1 as shown in Fig. 2. Additional accelerometers were installed at the center of front face of each 228
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 4. Shaking tank, pluviation device and the conveyor belt.
Table 6 Characteristics of shaking. Test no
1 2 3 4
Frequency (Hz)
3.0 7.0 4.0 4.0
Base acceleration (g) FHW
SHW
0.063 0.080 0.219 0.346
0.062 0.0827 0.216 0.349
Fig. 5. The base acceleration-time history for Test No: 3. Fig. 7. Total earth pressure and its two components for the bottom and top blocks of the FHW, Test No: 3 (a) Block 1, (b) Block 6.
Fig. 6. Block amplifications for the harmonic motions of (a) FHW and (b) SHW.
accelerations were observed to be linearly amplified when propagating from the bottom to top for maximum base acceleration amplitudes less than 0.2 g. For base acceleration exceeding 0.2 g, nonlinearly amplifying acceleration levels were observed and additionally the back-face
Fig. 8. Comparison between the oscillating earth pressures and the acceleration of the block for the FWH, Test No: 2, Block 3.
229
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 9. Distribution of the total earth pressure with depth for aavg,b < 0.2 g (Test No: 2).
Fig. 11. Pore water pressure ratios for FHW, (Test No: 3).
3.4. Displacement The displacement transducers were installed on each block of both model walls to determine sliding and tilting responses of the walls as shown earlier in Fig. 2. The horizontal displacements of each block (DP1 through DP6) were monitored at the center of the blocks whereas the vertical displacements (DP7) were measured at the top of the walls. Despite firm bottom conditions, small tilt angles were measured towards the water side under strong shaking without sinking into the seabed (Fig. 12). A comparative evaluation indicated that the tilt angle of SHW was about 1.5 times larger than that of FHW at the largest base accelerations corresponding to test 4. Illustrative horizontal displacement - time histories of each block of FHW and SHW models are shown in Fig. 13a and b, respectively. As revealed by the figures, the horizontal displacements increase with time, during the test duration and also block displacements increased with time and elevation. The normalized horizontal displacements (Dx/ z) recorded at the 10th second of the shaking versus base acceleration pairs were given in Fig. 14 for each wall type. It should be noted that z is the elevation difference of the displacement transducers relative to the backfill surface. It was clearly observed that the relative displacements increased with increasing base acceleration. Moreover, Fig. 15 showed that, the relative horizontal displacement was larger, the extent of which reached to a maximum of 1.8 times for SHW as compared to the ones of FHW model. The differences in the responses were attributed to a down-drag forces and the weight of the soil block resting on the widest block. In simpler terms, during model preparation, the backfill soil settled under its own weight, and mobilized a down-drag force, which mobilized some friction between the wall and the backfill soil. This frictional force and the weight of the soil resting on the widest positive back-slope of FHW stabilized the wall and increased the resistance against overturning and sliding. Since, the frictional surfaces of each block of the hunchbacked walls are different, the normalized horizontal displacements with block breadths were also considered. The results were compared for both hunchbacked walls as shown in Fig. 16. The normalized displacements (Dx/B) were relatively larger for SHW. The normalized displacements of the widest block (i.e.: the broken point of the walls) were found to be relatively small when compared with the other block displacements for both models. The differences in normalized displacements between two model types reached to a maximum of 33% at the top block for Test No. 3.
Fig. 10. Distribution of the total earth pressure with depth for aavg,b > 0.2 g, (a) FHW and (b) SHW (Test No: 4).
shape of the wall especially for both hunchbacked walls became an important factor of the dynamic interaction between the backfill soil and the wall. As shown in Fig. 10, for base acceleration amplitudes exceeding 0.2 g, lateral earth pressures were observed to increase linearly till the widest block, then a sudden decrease was observed at depths immediately below the widest block. This sudden decrease in lateral earth pressure was attributed to shadowing effects of the widest block.
3.3. Pore water pressure To better present the pore pressure responses of the backfill soil, a dimensionless parameter: excess pore water pressure ratio, ru was defined as given in Eq. (4).
ru =
Δu ′ σv,0
(4)
where Δu is the excess pore pressure, and σ′v,0 is the initial vertical effective stress. When Δu becomes equal to the initial vertical effective stress, then instantaneously vertical effective stress becomes zero during cyclic loading leading to ru value of 100%. Since shear strength in cohesionless soils is proportional with normal effective stress, an ru value of 100% suggest very low shear strength values. This phenomenon is widely referred to as soil liquefaction. In this study, the pore pressure transducers, PP1 and PP2 were installed 20 cm apart from each other for both wall types as shown earlier in Fig. 2. Due to high hydraulic conductivity of the coarse backfill soil, the excess pore water pressures developed during shaking were observed to be negligibly small as shown in Fig. 11.
230
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
Fig. 12. Schematization of the tilting behavior of the (a) first and second-type hunchbacked walls and (b) their tilting responses, respectively.
Fig. 13. Horizontal displacement-time histories of both hunchbacked walls (Test No: 3).
Fig. 14. Relative horizontal displacements (t = 10 s) for (a) FHW and (b) SHW.
3.5. Settlement
the total depth of the backfill soil (HB), and these values vs. base accelerations were shown in Fig. 18. As revealed by the figure, the normalized settlements increased with increasing base acceleration amplitudes addressing the nonlinear response of backfill soil. Comparing the results for both hunchbacked walls, especially for larger base accelerations (aavg,b > 0.2 g), the normalized settlements were relatively smaller for FHW. This response is rather expected since smaller
The surface settlement profiles of the backfill were developed by monitoring the surface elevations before and after shaking. As shown in Fig. 17, it was observed that the maximum settlement (ZB) was observed to be not located immediately behind the wall but ¼ H away from the walls and the settlements for SHW were observed to be larger than that of FHW. The maximum settlement of each test was divided by 231
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
during and after shaking wall displacements, total earth and pore pressures were monitored. Additionally, after the end of shaking, the soil profiles behind the model walls were also monitored. Followings are the specific conclusions of these studies: 1. From inertial forces point of view, the FHW as compared to SHW showed a significantly better performance level and are subjected to less normalized inertial forces. 2. Consistently, the average amplification ratios recorded at the top block of SHW were observed to be approximately 75% larger than those of FHW. 3. The oscillating component of the earth pressure and accelerations of the blocks were observed to be in-phase, which is valid for all blocks of both of the hunchbacked walls. 4. The variation of lateral earth pressures with depth were observed to be linear for maximum base acceleration amplitudes less than 0.2 g. 5. While for base acceleration exceeding 0.2 g, nonlinearly amplifying acceleration levels were observed and the back-face shape of the wall (i.e.: the wall type) became an important factor of the dynamic interaction between the backfill soil and the wall. 6. The observation of horizontal displacement, tilting and settlement results revealed that the first type hunchbacked wall exhibited a more stable performance as compared to the performance of the second type hunchbacked wall, especially at the base acceleration levels exceeding 0.2 g. 7. Consistently, the maximum tilt angle and the normalized horizontal displacements of SHW were observed to be about 1.5 and 1.8 times larger than that of FHW, respectively. 8. The settlements were also observed to be larger for SHW.
Fig. 15. Comparison of the relative horizontal displacements (Dx/Z) for both hunchbacked walls (Test No: 3).
Fig. 16. Comparison of the normalized horizontal displacements (Dx/B) for both hunchbacked walls (Test No: 3).
As the concluding remark, the seismic performance of these hunchbacked walls is controlled by their different back-face geometries, and larger positive hunched walls (i.e.: FHW) perform significantly better under seismic excitation leading to more economical and safer design.
Fig. 17. Results of the profile measurements for both hunchbacked walls (Test No: 3).
Acknowledgement We would like to thank to Scientific and Technological Research Council of Turkey (TUBITAK) Grant No: 113M426. References [1] Alyami M, Rouainia M, Wilkinson SM. Numerical analysis of deformation behavior of quay walls under earthquake loading. J Soil Dyn Earthq Eng 2009;29:525–36. [2] Arablouei A, Gharabaghi ARM, Ghalandarzadeh A, Abedi K, Ishibashi I. Effects of seawater-structure-soil interaction on seismic performance of caisson type quay wall. Comput Struct 2011;89:2439–59. [3] Fujiwara T, Horikoshi K, Higuchi Y, Sueoka T. Estimation of dynamic displacement of gravity type quay walls based on centrifuge modelling, In: Proceedings of the 12th world conference on earthquake engineering, Japan. [4] Iai S. Similitude for shaking table tests on soil-structure-fluid in 1g gravitational field. Soils Found 1989;29(1):105–18. [5] Inoue E, Miura K, Otsuka N, Yoshida N, Sasajima T. Numerical analysis of the earth pressure during earthquake on the gravity type quay wall, In: Proceedings of the thirtheenth (2003) international offshore and polar engineering conference, 2003, 25–30 May 2003, Hawai. [6] Kim SR, Jang IS, Chung CK, Kim MM. Evaluation of seismic displacements of quay walls. J Soil Dyn Earthq Eng 2005;25:451–9. [7] Kim SR, Kwon OS, Kim MM. Evaluation of force components acting on gravity type quay walls during earthquakes. J Soil Dyn Earthq Eng 2004;24:853–66. [8] Mendez BC, Botero E, Romo MP. A new friction law for sliding rigid blocks under cyclic loading. Soil Dyn Earthq Eng 2008;29:874–82. [9] Miura K, Kohama E, Inoue K, Otsuka N, Sasajima T, Hayashi T, Yoshida N. Behaviour of gravity type quay wall during earthquake regarding dynamic interaction between qaisson and backfill during liquefaction, In: Proceedings of the 12th world conference on earthquake engineering, Japan. [10] Mohajeri M, Ichii K, Tamura T. Modification of the sliding block concept for caisson walls. J Waterw, Port, Coast Ocean Eng, ASCE 2002;134:134–42. [11] Nakahara T, Kohama E, Sugano T. Model shake table test on the seismic performance of gravity type quay wall with different foundation ground properties, In: Proceedings of the 13th world conference on earthquake engineering (13WCEE),
Fig. 18. Normalized settlements with respect to base accelerations for both hunchbacked walls.
horizontal displacements of the wall lead to reduced settlements behind the wall. Also, the larger positive slope of the wall was believed to stabilize the wall subjected to shaking. On the basis of this, it can be concluded that first type hunchbacked wall has a zone immediately behind the wall, where it is safer to locate industrial and coastal structures. This observation is of value to reduce the damages to cranes and pavements, if they are located in this less settled zone behind the quay walls. 4. Summary and conclusions Within the confines of this manuscript, seismic responses of two different hunchbacked block type quay walls with different back face shape were presented. A series of 1-g shaking tank tests was performed using a 1/10 scaled block type quay wall with gravel backfill materials on firm non-liquefiable sea bed conditions. Sinusoidal harmonic shakings with frequencies ranging from 3 Hz to 7 Hz, and average base accelerations in the range of 0.062 g to 0.349 g were applied. Before, 232
Soil Dynamics and Earthquake Engineering 101 (2017) 225–233
Z.T. Yuksel et al.
[15] Tiznado JC, Rodriguez-Roa F. Seismic lateral movement prediction for gravity retaining walls on granular soils. Soil Dyn Earthq Eng 2011;31:391–400. [16] Towhata I, Alam MJ, Honda T, Tamate S. Model tests on behaviour of gravity type quay walls subjected to strong shaking. Bull NZ Soc Earthq Eng 2009;42(1):47–56. [17] Yuksel Y, Alpar B, Yalciner AC, Cevik E, Ozguven O, Celikoglu Y. Effects of Eastern Marmara earthquake on the marine structures and coastal areas. J Water Marit Eng, ICE 2002;156:147–63. [18] Yuksel Y, Yuksel ZT, Cevik E, Orhan K, Berilgen M. Evaluation of the seismic performance of a caisson and L-type quay wall. Soil Dyn Earthq Eng 2017;92:537–50.
2004, 1–6 August 2004, Canada. [12] Sadrekarimi A. Seismic displacement of broken-back gravity quay walls. J Waterw, Port, Coast Ocean Eng, ASCE 2011;137/2:75–84. [13] Sadrekarimi A, Ghalandarzadeh A, Sadrekarimi J. Static and dynamic behaviour of hunchbacked gravity quay walls. Soil Dyn Earthq Eng 2008;28:99–117. [14] Sumer M, Ansal A, Cetin KO, Damgaard J, Gunbak AR, Hansen NEO, Sawicki A, Synolakis CE, Yalciner AC, Yuksel Y, Zen K. Earthquake-induced liquefaction around marine structures. J Waterw, Port, Coast Ocean Eng, ASCE 2007;133(1):55–82.
233