Soil Zonation and The Shaking Table Test of The Embankment on Clayey Soil

The main objective of this research was to model the zonation of wrap faced embankment on soft clay foundation, by applying a shake table test. Also, to investigate the dynamic behaviors of clay soil, such as acceleration amplification, displacement and pore water pressure of wrap faced embankment. This was done with respect to changes in frequencies of 1 Hz, 3 Hz, 5 Hz, 10 Hz, 12 Hz and 15 Hz respectively. Constant acceleration (0.1 g) and surcharge (19 Kg) were applied by using a laminar box, placed on a shake table testing machine. The main elements of this research were the laboratory test, which was used for preparing reconstitute soil samples, and wrap faced embankment with frequency arrangement. After applying all test parameters, dynamic parameters were increased by rise in elevation with respect to frequency. The result shows that the maximum dynamic parameters were found at the frequency of 10 Hz. It is beneficial to the relative performances of the wrap faced embankment, which is the updated design parameter.


Introduction
The soil-foundation formed from soft clay becomes the focus of Seismic Engineering.
In some cases, the foundation on soft clay is creating a problem for the design and construction of any type of structure. In Bangladesh, the southern part of the country, an excessive amount of soft clay is found in the Khulna and Bagerhat districts and also in the surroundings of Dhaka city (Hore et al., 2019). Bangladesh has the largest delta in the world. Therefore, very large alluviums are deposited on its surface. The oldest deposits are the barend, madhapur and lamaicregion clay. The sediments deposited are not evenly distributed throughout the country. At the northern part, it is about 128 m thick and this is where granite is extracted for construction purposes. Conversely, the thickness was gradually increased towards the south. At the centre, the capital city, Dhaka of Bangladesh has a sediment covering of over 22 km (Al zaman & Jahan Monira, 2017; Alam & Islam, 2009;Bazlar Rashid et al., 2018;Haque et al., 2013;Hore et al., 2019). Earthquakes or any seismic effects may cause a major damage on this type of soil. The earthquakes or any sinusoidal waves (railway vibration) have often created significant problems during the design and construction of the embankment. Moreover, it may be damaged due to the softening of soilfoundation (Ering & Sivakumar Babu, 2020;Krishna & Latha, 2007;Kumar et al., 2020;Latha & Krishna, 2006;Zhou et al., 2020). The Soil and Structure Interaction (SSI) system was considered in this experimental study, to simulate actual soil-foundation (Bullock et al., 2019;Çelebi et al., 2019;He & Jiang, 2019;Hore et al., 2020;Srilatha, et al., 2013;Srinivasan et al., 2016). Soft soils in Dhaka, Bangladesh were used to build the soilfoundation. Krishna & Latha (2007) determined the result of shake table tests on geotextilereinforced wrap faced soil-retaining walls. A total number of 9 model tests were described, such as development, experiment methodology, and outcomes. Srilatha et al., (2013) illustrated the effect of frequency of base shakes on the dynamic response of unreinforced and reinforced soil slopes. Xiao et al., (2014) investigated the earthquake response of a slurry wall and presented a minimized scale shake table test. In this research, soil-cement-bentonite (SCB) was evaluated. This is also a common type of slurry wall. Reinforced soil is only slightly damaged during the recent seismic disasters in Japan (Suzuki et al., 2015). The seismic behaviour of this soil wall is formed from clay, cement-treated clay, sand based on shake table tests and results of the pull-out. Fleming et al., (2016) conducted a contemporary phenomenon on soft clay soil, to determine the consequences of soil improvement on the seismic resistance of piles. Yazdandoust (2017) conducted a recent laboratory test to assess the character of 1-g shake table tests. Many researchers also performed a shake table test to study the seismic responses of different soil structures on soft soil (Beskhyroun et al., 2011;Ripon Hore et al / GEOSI Vol 5 No 2 (2020) 196-209 Edinçliler & Toksoy, 2017Hamayoon et al., 2016;Hassan & Pal, 2018;Helwany et al., 2017;Madhavi Latha & Manju, 2016).
Few researchers focused on shake table tests on soft clay soil in South Asian regions.
In Bangladesh, there have been limited studies related to shake table tests of wrap faced retaining walls, on soft clay. Therefore, the purpose of this research was to study the effects of frequency on reinforced soil wall models. The objectives of this study were to: a) measure the response of wrap faced and reinforced soil retaining wall, which was subjected to dynamic loading through shake table; b) investigate the acceleration amplification, deformation, and pore water pressure with respect to frequency response; c) to zone the wrap faced embankment on the soft clay foundation, by applying a shake table test, and d) draw the contour maps using PLAXIS 3D Software.

Area of Study
The clay soil sample was collected at a depth of 1.5 m below the existing ground level from the Dhaka city of Bangladesh as shown in figure 1. These homogeneous, stiff, reddish brown samples were at first oven-dried. Subsequently, the dry lumps were then powdered gently by using a wooden hammer. It was finally sieved through a 200 standard sieve to obtain clean clay-like soil powder. This type of clay is very dominant in catchment areas (areas around ariver) with a thickness from 1 to 20 m. A total number of 1000 drill holes were carried out for the Standard Penetration Test (SPT) in and around Bangladesh.
Furthermore, 470 drill holes were chosen for the formation of soft clay layers.
In addition, a map identified with zoning for SPT values were from 1 to 5. The total area was divided into five subsections according to the thickness of soft clay soil. The lower range of soft clay thickness is 0 to 1 and the higher range is between 10 to 20. From the thickness map and bore log, SPT N zone was realized. The target of this study was to obtain the accuracy of the seismic design of roadway and rail sub structure, based on the aforementioned necessities. On this model embankment, 100 shake table tests were experimented. The analysis implemented a repeated loading and unloading process.

Using Equipment and Materials
The shake table facility, which is computer-controlled, was used to simulate the horizontal shake action associated with dynamics. The platform of testing was a square, with a dimension of 2.5×2.5 m² and an approximate payload capacity of 1100 kg, made up of steel plates. The range of acceleration is 0.05 g to 2 g. The frequency range is 0.05 to 50 Hz with a maximum amplitude of ±200 mm. The highest velocity is 0.03 m/s. The shake table test machine is shown in Figure 2. The laminar box is positioned on the shake table as presented in Figure 3. Every layer consists of an internal structure with internal dimensions of 915 mm × 1220 mm × 1220 mm.
A Laminar box was constructed on the shake table apparatus to reduce boundary effects wherever possible. The Laminar box neither resists nor promotes soil displacement to accommodate the movement of soil. The laminar shear box has 24 hollow aluminum layers of frames developed at BUET. Every layer consists of an internal structure with internal dimensions of 915 mm × 1220 mm × 1220 mm. Sylhet sand was used as the backfill material that was available locally. The unified soil classification system classified the sand as poorly graded sand (S.P). The maximum and minimum dry densities were 18 KN/m³ and 16 KN/m³m3, respectively. The specific gravity of the sand particles was 2.34. The relative density of sand was 60%. The soft clay soil in Dhaka was used in this research work.The liquid limit and water content of this soil sample were found at 40% and 23% respectively.
The soil sample used was prepared, by using 50% of water content (1.25 times of liquid limit). Cohesion was obtained by 14.8 KN/m² and friction was obtained by 1.0 from the direct shear test. The water content of the soil sample was 14%, and the unconfined compressive strength (qu) was 19 kPa, after the loading had been done. The thickness of the clay layer in the soil sample was 6 m. Reconstituted soil sample preparation was displayed in Figure 4. Figure 5 illustrates the schematic diagram of the test configuration. The dimension of the laminar box was 1220 mm deep, and the size of the plan area was 915 mm × 1220 mm. In lifts of equal height, the model was constructed while reinforcing each lift with a layer of woven geotextile. To measure pore water pressure, two pore water sensors (P1 and P2) were placed as shown in Figure 5. Besides this, 6 acceleration sensors (A1, A2, A3, A4, A5, and A6) were placed on different points of the model, to measure acceleration. A1 and A2 sensors were placed in clay soil layers. Three displacement sensors (LVDT1, LVDT2, and LVDT3), were placed on different locations of the model to measure the displacement of the embankment. Table 1 shows the test sequence.

Results and Discussion
The total number of 100 shaking A total number of 470 layers of soft clay soil were also presented in the GIS interface map. The Standard Penetration Test (SPT) zonation map was prepared based on SPT N value.
The SPT zone map has four sub sections. The lower range of SPT zonation map is 1 to 2 and the higher range is greater than 5. The N value is less than five, showing the existence of soft clay content of that area. This influences the dynamic behaviours of the embankment, (Hore et al., 2019). From the map ( figure 6), the green colour shows the SPT value between 1 to 2.
The capital of Bangladesh, Dhaka, is occupied with soft clay zone containing SPT value 1 to 2. The output result was the dynamic behaviour like acceleration amplification, displacement and pore water pressure. This was shown in the contour map and the graph. The research Ripon Hore et al / GEOSI Vol 5 No 2 (2020) [196][197][198][199][200][201][202][203][204][205][206][207][208][209] showed that dynamic behaviour of the soil depends on the soil standard penetration test result. The zonation map using Standard Penetration Test (N) value was shown in Figure 6. were amplified closer or slightly more than 1 for the frequency of 1 Hz. The differences in acceleration amplification for various frequencies were increased with increase in wall height. At a normalized height of 1, for 1 Hz, 3 Hz, 5 Hz, 10 Hz, 12 Hz, and 15 Hz, the values for acceleration amplification were 1.04, 1.24, 1.46, 3.26, 2.47, and 1.80 respectively.
These test results of acceleration are partially similar to the (Cai et al., 2019;Fleming et al., 2016;Hore et al., 2020). Figure 9 shows the displacement profiles observed for tests FT1, FT2, FT3, FT4, FT5, and FT6 with frequencies 1 Hz, 3 Hz, 5 Hz, 10 Hz, 12 Hz and 15 Hz respectively. It was observed that at the highest elevation (z/H=0.875), the displacement was maximum. The highest normalized displacement of 2.01% was observed for 12 Hz frequency. The corresponding values for the frequencies of magnitudes 1 Hz, 3 Hz, 5 Hz, 10 Hz and 15 Hz were 0.02 %, 1.99 %, 2.05 %, 2.01 % and 1.93 % respectively. These test results of face displacement response are partially similar to the (Krishna & Latha, 2007;Srinivasan et al., 2016;Suzuki et al., 2015).    Figure 11 shows the contour map (PLAXIS 3D output results) of acceleration, displacement, and pore water pressure response.    4 m and therefore, the model was developed. The length (L) of the geotextile reinforcement was 3.75 m. Also, 20 cycles of sinusoidal shakes were subjected to the model wall. All of the present model walls were later constructed with sand placed on the same medium to loosen density. The sand embankment properties were as follows: i) average unit weight ii) the relative density which were 18 KN/m3 and 60% respectively. These results obtained were similar to that of other researcher swho only used sand embankment without clay soil (Eric et al., 2013;Goktepe et al., 2019;Srilatha et al., 2013). The cyclic behaviour of the soft clay has been analysed based on the different input frequency.

Face Displacement Response
The face deformations were high for low-frequency shaking, low for surcharge pressures and high for base accelerations. The pore water pressures were observed to increase with a rise in base motion frequency. Three contour maps were drawn to investigate the acceleration amplification, deformation, and pore water pressure with respect to frequency response. These maps also showed dynamic parameters such as acceleration amplification, deformation, and pore water pressure representations of the soft soil. These results were very important for observing the dynamic behaviour of wrap faced soil retaining walls on the soft clay layer. This can be applied to improve incorporating dynamic loading, considering the design specification of this type of retaining wall (Railway and Road embankment).

Conclusion
The shake table test on wrap faced embankment on the clay soil foundation, is a new form of test. The lower range of SPT zonation map is 1 to 2 and the maximum range is greater than 5. In Bangladesh, embankment on soft clay soil plays a very vital role in seismic perspective. Therefore, the research opportunity of this area will bring about an upgrade to seismic design specifications. From the test results, it was discovered that face deformation increases with rise in elevation. At high elevations and frequencies, pore water pressure is also high. These test results are beneficial in understanding the relative performance of the wrap faced embankment, which is the indicative direction to the design process of embankment.