Journal of Engineering Geology

---

(Paper pages 523-542) This paper presents a rigid circular footing model with specified properties and dimensions on a sandy-clay soil with Mohr-Coulomb material. This model is analyzed dynamically with finite difference 2D FLAC software under vertical component of ground excitations. Then the soil is improved with cement grouting and analyzed again. Consequently, the load-settlement curves under a circular footing, due to vertical component of ground accelerations through the underlying soil, are plotted. Also the dynamic bearing capacity of natural and soil cemented foundation is presented and discussed. The analysis results show that adding 2, 4 and 6 percent of cement, with certain conditions, cause 2.7, 4.2 and 7.0 times increase in dynamic bearing capacity, respectively, in comparison to normal soil.

---

Construction of asphaltic core dams is a relatively novel method especially in Iran. Iran is located in a region with high seismicity risk. Therefore, many researchers have focused on the behavior of such types of dams under earthquake loading. In this research, the behavior of asphaltic core rockfill dams (ACRD) has been studied under earthquake loading using nonlinear dynamic analysis method and a new method is presented to assess seismic stability of these types of dams in earthquake conditions. Based on nonlinear dynamic analysis, the current study attempts to provide an appropriate criterion for predicting the behavior of earth and rockfill dams considering real behavior of materials together with actual records of earthquake loading. In this method, the maximum acceleration of the earthquake record (PGA) increases until instability conditions. Finally, a new criterion is presented for evaluating seismic safety of ACRDs via demonstrating curves of the crest's permanent settlement and maximum shear strain against maximum earthquake acceleration. Results of the proposed criteria can assist designers of asphaltic core dams to predict dam stability during earthquake event

---

This paper presents a case study of the instability mechanism, to verify and reinforcement method adopted construct collapsed zone of Sabzkuh water conveyance tunnel in southwest Iran. The instability problems were encountered during tunnel excavation due to the failure, changes in stress field lead to deformation causing dilation and increasing the permeability of sand and gravel layers, local fault gouge zones, landslide and in turn significant reduction in shear strength and collapse in tunnel. IPE Arch Support Technique (IAST) was, used for T1 part of Sabzkuh tunnel zone in order to reinforce the ground around tunnel and to cross the zone falling. In this study, Finite Element Method was employed for the quantitative reinforcement effect with deformation modulus of ground, IPE length and size. As a result, the settlement increases as length increases and decreases with the increase of the deformation modulus of ground and IPE size.  

---

One of the best approaches to reduce transportation problem is to use the underground tunnels. Therefore, Niayesh highway tunnel was performed by the New Austrian Tunnelling Method (NATM) in the northern part of Tehran and it includes north and south tunnels. The excavation of tunnels and other underground structures cause considerable changes in local stress conditions around structures that lead to surface settlement. In this research, surface settlement has been studied for five sections (CS-1 to CS-5) by empirical methods, numerical analysis and actual settlements. For the empirical and numerical methods, O’Reilly and New (1982) method and also finite element method (PLAXIS_2D software) have been used, respectively. On the basis of the obtained results, the numerical method in all sections (except section 3) is in agreement with the actual settlements. While, empirical methods have estimated the settlements more than actual values in those sections.  Also, the achieved results from the aforementioned methods show that the maximum settlement due to tunnel excavation is more than allowable settlement and it is in risk condition

---

In tunnelling in soil mass, in groundwater existing mode, liquefaction, elastic displacements and settlement in soils upon the tunnel, are the risks may attack the excavated underground space stability. In this case study that were performed on second line of Mashhad city subway route, information catched from Standard Penetration Test, in situ and laboratorial tests, were used to optimum numerical values search for soil engineering parameters that could optimize the TBM stationing level. In order to this goal attaining, intelligent, numerical and probabilistic methods were used and the reliability of intelligent and numerical methods with the Safety Factors of tunnel stability, investigated simultaneously. The results were denoting the accordance of intelligent models such as Genetic Algorithm (GA) and Multi objective Genetic Algorithm with Finite Element model's output. So these models could be complement of each others in planning and designing of tunnels and using of them advised in tunneling and excavations.

---

Ground settlement due to tunneling and the effects of the engineering geological factors on its dimensions and extensions, is a very important problem in shallow tunnel excavation projects in urban areas. Empirical method is one of the usual methods to study this subject. The empirical and dimension-less parameters of VL and k are the most important parameters in relation to this method that are estimated according to engineering geological factors. In this research, the values of these parameters were initially estimated based on preceding studies and the ground settlement was predicted using these estimated values of VL and k. In next stage, the results of predictions were compared with the real (measured) settlements happened due to Abuzar tunnel excavation. As the real settlements are less than the predicted ones, it was concluded that the real VL must be lower than the predicted values or the real k must be higher than the predicted values. With regard to the high dependency of these parameters to the soil cohesion, it seems natural cementation of Tehran alluvia has acted as a factor to increase the soil cohesion and has caused to decrease ground settlement due to excavation of Abuzar tunnel. For validation of this hypothesis, preceding findings about alluvia cementation were reviewed and the results of in-situ and laboratory shear and triaxial tests were compared with together. Then it is concluded that the higher cohesions of in-situ shear tests are occurred due to natural cementation of materials existing in Abuzar tunnel route

---

./files/site1/files/2.pdfExtended Abstract
(Paper pages175-200)
Introduction
In weak soils with low bearing capacity, the load transfer is done using piles. Therefore, by creating an interposed layer separating the pile from the raft, reactions between raft and pile head will be reduced and the load-bearing role of shallow soil will be more than contact pile situation. Normally, the pile head and shallow soil have a settlement equal to the raft. Thus, the relative settlement of pile and soil in pile head is equal to zero and at the bottom is high and the body friction mobilizes upward. In addition, a portion of load is tolerated by shallow soil and the other portion is tolerated by the pile head, which would be transferred to deeper soil layers. In noncontact state, with the formation of a hard soil layer on which the raft is located, soil mechanical parameters will be improved; while in contact state, the settlement will be decreased by reducing the amount of transferred load to the shallow soil. The transferred load to the shallow soil increases vertical and horizontal stress around piles, so bearing capacity of piles is increased.
Methodology
In this study, a parametric study has been performed concerning contact and noncontact piles using finite element software namely, ABAQUS/CAE software version 6.13.1 and the obtained results were compared (with what? The sentence is incomplete). Thus, simulations are is done for states of 0, 1, 4 and 9 piles for each of the contact and noncontact piles (total of 8 simulations). In the present research two models were taken to investigate the optimum mesh sizes, 12 models for parametric studies on parameters of piles’ length, piles’ diameter, thickness of the raft and interposed layer and one model for verification study. Models in both contact and noncontact have been considered with a one meter interposed layer. Raft width and thickness were selected 7.5 and 1.6 m, respectively. Width and depth of the soil mass used in the model were 32 and 26 m, respectively, and the distance between the bottom of the pile and the soil mass was 13 m. In all cases, the diameter of piles was 0.5 m and distance between piles were 5 and 2.5 m in 4 and 9 states, respectively. The geotechnical parameters and model dimensions used, were selected according to the Fioravante & Girettis (2010) [1]. Sand and silica-sand with the defined properties were used for the soil mass and the interposed layer, respectively. Since Drucker-Prager criteria has better ability to express the behavior of coarse-grained soils, this criterion was used in the modeling [2]. The purpose of this study is to investigate the influence of interposed layer on bearing capacity and settlement of pile. Hence, because of simplifying the process of modeling, parameters of main soil and interposed layer are mostly similar. Piles and raft are made of concrete with an elasticity modulus of 21 GPa, Poisson's ratio of 0.2 and density of 2300 kg/m³. The crack growth analysis with the compressive stress-plastic strain was used to express the fracture behavior of concrete [2, 3 & 4]. In the present study, frictional and vertical contacts between surfaces were considered for conducting interactions between different materials. For frictional contact, the penalty formulation with the fixed friction coefficient of tanδ was used where δ is the angle of friction. The penalty formulations and hard contact were applied between two surfaces for the normal contact. Interactions were considered in the modeling including raft-soil mass, raft-interposed layer, pile-raft, interposed layer-soil mass, interposed layer-pile and the soil-pile [5 & 6]. Coefficient of soil lateral pressure used in this study corresponds to k₀=0.65 which is introduced in many geotechnical conditions [7]. A uniform distributed vertical load 500 kPa was applied on the raft. For getting results in every portion of loading time, this amount is applied in order of 5 kPa in each time interval. To accelerate the process of analysis and because of the symmetry of all models in two directions of X and Y, the quarter model technique was used, so that movements in the direction perpendicular to the sheet and rotation around perpendicular axes on the sheet were not allowed on the border of symmetry. The boundaries of the models due to the enough distance from the piles were considered in a way that lateral displacement and rotation around the vertical axis was not allowed. Furthermore, the bottom of the soil mass was considered as complete fix due to the enough distance from the pile foot.
Conclusion
In this research, a numerical – parametric study is performed on special kind of piles named noncontact piles and results are compared with contact piles. Results of this study can be summarized as follows:
1. By increasing the number of piles from 1 to 9, the settlement reduced more in a noncontact state showing more effectiveness of implementing 9 contactpiles and thus requiring more piles in this case.
2. Soil surface stress differences in noncontacts states from 4 to 9 piles was less than contact state (approximately 1/7) indicating that more piles is needed to conduct the contact state.
3. Stress changes in the soil under the pile in noncontact state by adding piles from 1 to 4 was higher than adding piles from 4 to 9 indicating the suitability of using 4 noncontact piles; while, in the contact state, the stress changes in the soil under the pile in both cases from 1 to 4 piles and from 4 to 9 piles was noteworthy showing the necessity of using the ninth pile.
4. Unlike the states of 4 and 9 piles, the negative friction in noncontact state and 1 pile was seen along the piles, which can be due to the fewer piles and the effect of interposed layer density as well as soil mass at greater depthsbecause of lesser effect of piles in load-bearing.
5. The ratio of heads load in the contact to the noncontact piles was about 2.5 to 4 reflecting the positive impact of using interposed layer on load reduction and smaller cross-layer design for piles. In addition, the ratio of heads load in the contact to the noncontact piles was higher for 4 piles than 9 piles that represented the suitability of using 4 piles.
6. Based on the results of geometric parametric studies it is found that:
(A) By resizing the elements from 0.25 to 0.5 m, the results had not changed and only time of analysis was increased.
(B) Among three values of 0.5, 1 and 1.5 m for interposed layer thicknesses, the thickness of 1 m was enough and had a good effect on the stress distribution and involving shallow soil in bearing vertical stress.
(C) The raft thickness of 1.6 m was appropriate so that with this thickness, the resultant effect of increasing vertical loads (raft weight) and increased rigidity due to increased raft thickness caused the stress and settlements remain in a reasonable range.
(D) Due to the increased friction by increasing in diameter, the optimal diameter of 0.5 m was achieved for piles which reduced the settlement by receiving more load.
(E) Among three pile lengths of 10, 19 and 25 m, the optimal length was 19 m; so that by further increase in the length, stresses and settlements were not noticeably changed.In total, noncontact piles had better performance compared to contact piles in similar conditions.
Reference
1. Fioravante V., Giretti D., "Contact versus noncontact piled raft foundations", Can. Geotech. J. 47 (2010) 1271-1287.
2. Saba H., "Verification of nonlinear condition of anchored walls in various loading", Thesis document of Amirkabir University of Tehran, Iran (2003).
3. Fioravante V., "Load transfer from a raft to a pile with an interposed layer", Geotechnique 61, No. 2 (2011) 121-132.
4. Dastani H., Shariati M., "Numerical and experimental analysis of controlling of crack propagation route in a plane under cyclic uniaxial loading by creating openness", Thesis document of Shahrood Industrial University of Shahrood, Iran (2014).
5. Randolph M. F., Wroth C. P., "Application of the failure state in undrained simple shear to the shaft capacity of driven piles", Geotechnique, Vol. 31, 1 (1981) 143-157.
6. Poulos H. G., Small J. C., Ta L. D., Sinha J., Chen L., "Comparison of some methods for analysis of piled rafts", Proc. 14th Int. Conf. Soil Mech. Found. Engng, Hamburg, Balkema, Rotterdam, Vol. 2 (1997) 1119-1124.
7. mottaghi A., "3D static and dynamic analysis of pile group with considering soil-pile interaction", 6th National Congress of Civil Engineering, Iran, Semnan (2012).

---

./files/site1/files/5.pdfExtended Abstract
(Paper pages 247-276)
Intro

---

Introduction
When loose sand is subjected to seismic shaking, it tends to volume reduction and settlement. The density of the under layers is revealed in the settlement of the ground surface that causes the destruction of the structures located on ground surface. In dry sand layer, settlement in severe shakings occurs under a constant and effective stress condition and very rapid stress. In this regard, the sand deposit settlement is completed before the end of an earthquake, but if the sandy soil layer is saturated and drainage is limited the condition is prepared of fixed volume situation and the major effect of the seismic shocks is generation of exceed pore water pressure. Therefore, the deposit settlement of saturated sand requires a longer time, varying from a few minutes to a few days, depending on the permeability and compressibility of the soil and the length of the drainage path. The main purpose of the present study is to evaluate liquefaction hazards along Tabriz Metro Line 2 with using Standard Penetration test (SPT) and shear wave velocity (Vs) methods. Also, the probable rate of settlement in the soil layers in study area and correlation with liquefaction potential index (LPI) according to both procedure have been determined and discussed in the following paragraphs.
Material and Methods
In order to evaluate the liquefaction potential of soils using two field methods, geotechnical information of 54 boreholes in Tabriz Metro Line 2 were collected. The types of soil and geotechnical properties can affect the liquefaction potential. In this study, the gravely sand, silty sand, silty and sandy soils were studied. Ground water level is one of the main parameters in in soil liquefaction potential evaluation of soils. Variation of water level in boreholes is 2 to 30 meters. The peak ground acceleration (PGA) is necessary for the analysis of boreholes to evaluate liquefaction potential of soils. PGA values were selected in each boreholes position according to the Iranian Code of Practice for Seismic Resistant Design of Buildings (Code-2800-ver.4) equal to 0.35g (for return period 475 years and design life 50 years). Liquefaction potential of soil layers based on SPT results with appliying Idriss and Boulanger (2010) method has been assessed. Andrus, Stokoe and Jung (2004) procedure was used in shear wave velocity (Vs) method (with assuming cementation and un cementation condition in soils). Liquefaction potential index (LPI) of soil layers was calculated for both field tests results. Then, probable rate of settlement due to liquefaction in saturate soil layers was determined. Tokimatsu and Seed (1978) method applied for SPT results, Yoshimine (1992), Yoshimine et al. (2006) and Yi (2009) procedures have been used in Vs test. Finally, correlation between rate of settlement and LPI results were determined.
Results and discussions
Outcomes of this study can be explained in below:
1. Results obtained from comparison of both methods in liquefaction potential evaluation have been showed, agreement between two methods have been happened rarely. Specially, with assuming cementation condition in soils, LPI obtained from Vs method is more than SPT. Although, different factors can be affected at uncertainties in SPT results such as type of drilling machine, energy efficiency and accuracy of test performing. Also, in shear wave velocity method, maximum velocity for occurring liquefaction in soil layers () related to fines content percentage. It is possible that boundary values in procedure not compatible with geotechnical properties in study area.
2. Evaluation of probable rate of settlements in soil layer in study area have been showed that settlement values obtained from Vs is more than SPT. This condition is compatible with LPI amounts.
Conclusions
In sum up, settlement due to liquefaction in saturate soil layers is one of the important phenomena in geotechnical earthquake engineering. Maximum rate of settlement in soil layers in study area is equal 0.45m based on SPT method and 0.9m according to Vs procedure which should be considered. Accordingly, serious damages can be inflicted to buildings, underground structures and life lines in study area.  Therefore, it is suggested in future researches with using empirical and numerical (or soft computing) methods based on field and experimental tests results a detailed assessment conducted and influence of various parameters on settlement of soil layers be determined and the items listed below should be considered:
- Cementation parameter (C) values of soils in shear wave velocity method maybe not compatible with geotechnical properties in study area. It should be evaluated exactly. 
- In this research, peak ground acceleration (PGA) value was selected based on code 2800-ver.5. As regard to Tabriz Metro Line 2 is beside to Tabriz North Fault, PGA value according to historical earthquake catalogue and seismic risk analysis should be evaluated and seismic hazard have to determine with accuracy. 

---

Introduction
The design engineers usually follow a specific decision-making process for optimal selection of the type of required foundation and its design. In this state, in case the surface foundation is not appropriate for the project conditions, before making any decisions about the use of deep foundations, the proper methods for optimization of the liquefied soil should be evaluated in order to compare the advantages and disadvantages of each of them with those of deep foundation, in terms of efficiency, implementation problems, costs, and finally to select the best choice. One of the best methods of soil improvement is the use of stone columns. The rationale behind the use of stone columns is the high shear strength of materials and the provision of lateral grip by surrounding soil. Therefore, the stone column can receive the load from the structure, and transfer it to the resistant layers. In the soils with low shear resistance, the lateral constraint crated by the surrounding soils is not enough for preventing the sideway buckling of the column under which is subjected to the loads. Thus, special measures should be considered for the use of stone columns in these soils. One of these methods is the use of reinforcement shelves such as geogrid and geotextile. Investigating the previous studies, the lack of evaluation of the design parameters such as the settlement ratio of the soil improved by the reinforced stone column to the geogrid, and provision of design graphs in this regard, has been revealed. Therefore, by extension of the studies conducted by Chub Basti et al. in 2011, the design graphs were provided in this regard.
Material and methods
The PLAXIS V8 Software was used for modelling the soft soil improved by the stone column. For increasing the precision of the results, the 15-knot element was used in the current study. The fine mesh was used in the models made for the analysis of the problem. For simulation of the improved soft soil with the stone column in a single cell, the modelling was implemented in a two-dimensional environment in axial symmetry conditions. In the current study, it was assumed the rigid foundation is on the improved bed. Thus, for analysis of the simulated model, a vertical strain up to 2% of the soft soil height has been applied on the ground. Also, for simulation of the soil behavior, an appropriate model of soil and parameters proportional to the materials should be allocated to the construct geometry. The non-linear stress-strain of the soil in different levels of the problem can be simulated. The number of model parameters increases with the level of problem rupture. For precise simulation, we need the proper parameters of the materials. For modeling of soft soils and stone columns, elastic-plastic model with Mohr-Coulomb rupture criterion was used. In the current study, it was assumed the soft bed is located on a very hard layer of soil. Therefore, the vertical deformation was prevented on this horizontal boundary. Also, the horizontal deformation in two vertical edges was prevented and only deformation in vertical direction was allowed. The soft bed close to saturation was considered without the determined free water level. For models with stone columns, the element of interface between the stone column and soft soil, has been used. The reason behind using this element is that the stone column rupture is of shear form and due to this, a significant shear stress is created on the common surface between the stone column and soft soil. The percentage of the replacement area is defined as the ratio of the total area of the stone columns to the total area of the non-improved area. In the current study, the percentage of the replacement area is utilized between 10 to 30%, which is used in implementation. Also, the diameter of the stone columns is from 0.6 to 1.2, in the analyses.
Results and discussion
The results of the numerical study were compared with the existing theoretical relationships provided by Poorooshasb and Meyerhof (1997), and Pulko et al. (2011). Figure 1 shows the comparison of the replacement percentage (RP) and settlement ratio (SR) in the non-reinforced state in the current study as well as theoretical relationships proposed by the previous researchers. Based on this figure, there is a difference between the results of the current study and those of Poorooshasb and Meyerhof (1997), and Pulko et al (2011). Poorooshasb and Meyerhof (1997) calculated the settlement ratio in their proposed material with the assumption of linear elasticity of the materials without consideration for plastic settlement. Therefore, the settlement of the improved soft soil with stone column, calculated by Poorooshasb and Meyerhof, would not show the real amount. However, Pulko et al. (2011), with consideration for the elastoplastic behavior of the materials, the lateral expansion of the stone column, and the primary stress of the soil around the column, provided more realistic results that correspond closely with the present study. Also, for designing the stone column, the results of its reinforcement have been also provided in the graph presented in Figure 2. Thus, by the use of these graphs, the ratio of settlement reduction can be obtained for each distance between columns and with different percentages of alternatives../files/site1/files/124/2jalili%DA%86%DA%A9%DB%8C%D8%AF%D9%87.pdf

---

Maximum surface settlement (MSS) is an important parameter for the design and operation of earth pressure balance (EPB) shields that should determine before operate tunneling. Artificial intelligence (AI) methods are accepted as a technology that offers an alternative way to tackle highly complex problems that can’t be modeled in mathematics. They can learn from examples and they are able to handle incomplete data and noisy. The adaptive network–based fuzzy inference system (ANFIS) and hybrid artificial neural network (ANN) with biogeography-based optimization algorithm (ANN-BBO) are kinds of AI systems that were used in this study to build a prediction model for the MSS caused by EPB shield tunneling. Two ANFIS models were implemented, ANFIS-subtractive clustering method (ANFIS-SCM) and ANFIS-fuzzy c–means clustering method (ANFIS-FCM). The estimation abilities offered using three models were presented by using field data of achieved from Bangkok Subway Project in Thailand. In these models, depth, distance from shaft, ground water level from tunnel invert, average face pressure, average penetrate rate, pitching angle, tail void grouting pressure and percent tail void grout filling were utilized as the input parameters, while the MSS was the output parameter. To compare the performance of models for MSS prediction, the coefficient of correlation (R²) and mean square error (MSE) of the models were calculated, indicating the good performance of the ANFIS-SCM model.

---

In the mechanized boring method, the factors affecting ground surface settlement can be mainly divided into five categories: geometric, geomechanic, boring machines working, operating and management parameters. In urban tunnels bored mainly in shallow soil bed, face pressure can be one of the factors preventing ground settlement. The Line A tunnel in Qom metro project is bored with an EPB (Earth Balance Pressure) mechanized boring machine. The effect of face pressure on ground surface settlement was analyzed in the present study according to five sections of the tunnel. These five sections were selected in different kilometers of the tunnel where settlement gauges were installed and the results could be validated. To investigate the effect of face pressure on maximum ground surface settlement, four pressure levels of 100 kPa, 150 kPa, 200 kPa, and 400 kPa were taken into consideration. These were 1, 1.5, 2, and 4 times of the initial face pressure level, respectively. The ground surface settlement was assessed at four pressure levels using the finite element software, PLAXIS 3D TUNNEL. The results were validated using ground-level instrumentation (settlement gauges) on all sections. The validation showed that the modeling results are in good agreement with the results obtained from settlement gauges.  Comparison of the results indicated that a 4-fold increase in the face pressure led to a maximum decrease of 4.45 mm in the maximum settlement. Therefore, an increase in the face pressure can reduce settlement, although quite minimally. It was also found that an over-increased face pressure (face pressure over 200kPa) not only did not reduce the maximum ground surface settlement but also may lead to passive failure or uplift of ground surface ahead of the shield. 
 

---

In urban areas, it is essential to protect the existing adjacent structures and underground facilities from the damage due to tunneling. In order to minimize the risk, a tunnel engineer needs to be able to make reliable prediction of ground deformations induced by tunneling. Numerous investigations have been conducted in recent years to predict the settlement associated with tunneling; the selection of appropriate method depends on the complexity of the problems. This research intends to develop a method based on Artificial Neural Network (ANN) for the prediction of tunnelling-induced surface settlement. Surface settlements above a tunnel due to tunnel construction are predicted with the help of input variables that have direct physical significance. The data used in running the network models have been obtained from line 2 of Mashhad subway tunnel project. In order to predict the tunnelling-induced surface settlement, a Multi-Layer Perceptron (MLP) analysis is used. A three-layer, feed-forward, back-propagation neural network, with a topology of 7-24-1 was found to be optimum. For optimum ANN architecture, the correlation factor and the minimum of Mean Squared Error are 0.963 and 2.41E-04, respectively. The results showed that an appropriately trained neural network could reliably predict tunnelling-induced surface settlement.

---

Introduction
In some ports, the dredging and accumulation of a large amount of sedimentary material turned to a serious challenge, because of their sequent environmental and economic effects. These problems clarify the necessity of reusing dredged materials. Often, owing to their poor mechanical properties, they are not applied directly in technically engineering uses, so they require to be improved. Geocell application is one of the methods used for the improvement of soil behavior, which confines the sand mass through itself in the three-dimensional structure. These methods ease the speed of applying emerged it into a perfect option for stabilizing of the granular soil.
 In Shahid Rajaee port, by the dredging process for developing new phases, a large amount of calcareous sand is being accumulated near the Persian Gulf coastline. Therefore, in order to provide a solution to reuse these materials, this study attempts to investigate the beneficial influence of reinforcing sand by geocell on its load-beneficial behavior experimented by the plat loading test. For this purpose, a large scale model including circular foundation on reinforced and unreinforced sand has been employed under cyclic loading process.
Material and Methods
Soils
Two types of soils were used in this study. The first type was the sand derived from the dredging process of Shahid Rajaee port which has been used in different layers of the models. The second type of soil was well-graded gravel which has been used only in the cover layer.
Geocell
The geocell in this study were made of heat-bonded non-woven polypropylene geotextiles. Single cells were 110 mm long, 100 mm wide and 100 mm height.
Plate load test
In order to determine the bearing capacity of backfills, repeating plate load test was used with 150 mm diameter. Loading process included four stress levels (250, 500, 750 and 1000 kPa) consisting of 10 cycles each.
Test backfills
Four backfills was made by manually compacting the dredged sand, with tamper up to 350 mm in reinforced cases and 450 mm in unreinforced cases. Then geocells placed and dredged sand filled with accuracy in cells. Finally, a 50 mm thick sand or gravel cover layer, was placed. All lifts were compacted to 70% of relative density with 4% moisture content.
Results and Discussion
PLT results are summarized in Table 1. According to the results, only geocell reinforcement backfills can carry standard truck wheel load (550 kPa). Geocell can increase the ultimate strength of backfills with a sand cover layer by 70% (from 416 kPa to 725 kPa) while in backfill with a gravel cover layer showed 80% increase (from 520 kPa to 960 kPa) in ultimate strength. The gravel cover layer in unreinforced backfills increases the ultimate strength by 25 percent (from 416 kPa to 520 kPa).
Table 1. Results of PLT and performance ratings

Backfill name	UR-S	GR-S	UR-W	GR-W
Maximum stress (kPa)	416	725	520	960
Settlement at failure (mm)	4.6	9.0	15.5	14.9
Plastic settlement (mm)	3.5	7.0	12.5	12.0
Number of load cycles	10	20	20	30
Bearing capacity ratio (BCR)	1	1.74	1.25	2.32
Performance rating	4	2	3	1

Base on Table 1, bearing capacity ratio (BCR) has been increased up to 2.3 and has best when geocell reinforcement and gravel cover layer were used together. Geocell utilization as reinforcement for sand backfills, improves the stress-settlement behavior. Dredged sand can be used as backfill material for yards and access roads when reinforced with geocell and covered with a layer of well-graded gravel../files/site1/files/132/3Extended_Abstracts.pdf