Using EPS Buffers for Diaphragm Walls

Transcription

1 Using EPS Buffers for Diaphragm Walls Beshoy M. FAKHRY1, Salem A. AZZAM2, and Sherif S. ABDELSALAM1 1 Graduate Student, the British University in Egypt, beshoy115377@bue.edu.eg 2 Teaching Assistant, the British University in Egypt, salem.azzam@bue.edu.eg 3 Associate Professor, the British University in Egypt, sherif.abdelsalam@bue.edu.eg ABSTRACT Retaining walls are essential for deep excavation and other infrastructure projects with complex subsurface conditions. Design and construction of a lengthy walls is quite complex especially when the wall height exceeds 6 m, or when multi-anchors system is required to support diaphragm walls. In an attempt to reduce lateral earth pressures acting on flexible diaphragm walls, or embedded walls, a compressible inclusion such as expanded polystyrene Geofoam (EPS) is a durable material that can be used as a buffer between walls and soil backfills. In this paper, a numerical model was developed using finite element (FE) program PLAXIS 2D. The constitutive properties utilized in the model were characterization based on an extensive laboratory testing program, whereas various thicknesses of EPS were utilized and simulated in the FE analysis using the hardening soil constitutive model. The model results were verified by means of a physical prototype that was assembled to mimic the modeled wall types and the EPS buffers. From the main outcomes, the lateral pressure on flexible diaphragm walls can be significantly reduced by 37% using a relatively thin EPS buffer. Keywords: Geotechnical, Geosynthetics, Geofoam, Diaphragm walls, EPS buffer, Lateral pressure, modeling.

2 EPS BUFFERS FOR RETAINING WALLS There is a complex equilibrium in earth retaining systems between loads and resistances. Global and local safety factors (FS) must be achieved in design and construction stages, whereas these FS should cover external geotechnical stability and also the internal structural capacity of the system. For cantilever walls, the factors of safety consist of FS against sliding, overturning, and soil capacity; therefore the soil at the foundation must produce enough friction to prevent excessive sliding. The soil must also provide sufficient shear strength to support the bearing pressure, while the design geometry of the wall must provide sufficient FS to resist overturning. For diaphragm walls, the structural stability require sufficient resistance against straining actions due to lateral earth pressure, which mainly comes from the soil behind the diaphragm wall, while the overall geotechnical stability must be achieved by conducting a slop stability check. Diaphragm walls are a sophisticated type of retaining systems than cantilever and gravity walls, also they are considered as expensive ones compare to the simple retaining system. Many of the large infrastructure projects such as underground mutli-story garages, tunnels and metro stations, large pipe lines, require free length to retain the lateral earth pressure, and in most cases a diaphragm wall with anchors is required. The soil lateral pressure behind the diaphragm wall represent the main acting loads that lead sometimes to wall oversized dimensions, and may require single or multi- anchors, which makes this type of retaining walls even more expensive. According (Horvath, 1994), lateral pressure can be reduced by using expanded polystyrene (EPS) geofoam instead of soils which considered as a lightweight alternative for soil backfill, as the relation between lateral pressure and backfill weight is proportional. Therefore, Soil backfill behind cantilever walls can be replaced by a material such as foam to achieve static stability (Karpurapu and Bathurst, 1992; Horvath, 1994; Zarnani and Bathurst, 2008; Lutenegger and Ciufetti, 2009; Athanasopoulos et al., 2012; and Ertugrul and Trandafir, 2013). This approach has proven to achieve the desired safety by reducing the acting loads rather than increasing resistance, leading to a significant reduction in the wall dimensions, but is mainly proven for rigid cantilever walls (Abdelsalam and Azzam, 2016). For flexible cantilever walls, only limited research was conducted indicating that the lateral earth pressure can also be reduced. For instance Ertugrul and Trandafir (2013) conducted a twodimension (2D) plane-stain numerical analysis using program FLAC 2D v.6 (Itasca, 2008) to

3 investigate the effect of EPS inclusion behind flexible walls and indicated that an additional reduction in active lateral pressure was achieved due to the arching mechanism induced in the soil backfill behind EPS. Recently, AbdelSalam et al. (in press) have modeled flexible cantilever walls using the three-dimension (3D) finite element (FE) software package PLAXIS 3D v.1 Anniversary Edition (Brinkgreve et al., 2015) to identify expected amount and possible reasons of reduction in lateral pressure. Based on AbdelSalam et al. (in press), it was noticed that an arch shape was evident in the flow of the lateral stresses after using EPS buffer, which was justified due to the relative large compression that occurred within the bottom segment of the EPS, cased part of the lateral stresses was carried by friction with the neighboring soils. Diaphragm walls can be considered as flexible embedded wall types, hence a reduction in the lateral earth pressure can be achieved behind diaphragms as well, but if an EPS sheet is placed as buffer between the wall and soil behind it. However, most of the current research focused on modeling EPS behind rigid walls, some recent research focused on flexible walls, and the rest focused on seismic behavior of EPS behind walls such as researches done by Zarnani and Bathurst (2007), Athanasopoulos et al., (2012), Padade and Mandal (2014), Abdelsalam and Azzam (2016), and others. There is almost no research that contain information from verified laboratory and/or numerical models to solve diaphragm walls with EPS buffer, and this could be due to the following: 1) method of placing EPS sheets between wall and soil was not simple; and 2) diaphragm walls are flexible and the behavior was not very well understood. In this paper, the EPS geofoam was proposed as a thin buffer between the diaphragm wall concrete section and the soil behind the wall, in an attempt to understand the change in the lateral pressure acting on the wall. A method of statement for placing the EPS buffer was proposed. Local EPS properties were summarized based on Abdelsalam et al. (2015) who conducted a series of laboratory tests for EPS characterization. Then, a diaphragm wall was modeled using the twodimension (2D) finite element (FE) software package PLAXIS 2D v.8.6 professional package (Brinkgreve et al., 2006) to identify the expected amount of reduction in lateral pressure behind such flexible wall type, and using various thickness of EPS buffers. Finally, the FE results were verified against measurements from a physical prototype constructed as part of this study. The main outcomes included design charts that show the expected overall reduction in the lateral pressure behind diaphragm walls with respect to the EPS thickness.

4 EPS MECHANICAL PROPERTIES Abdelsalam et al. (2015) have conducted a series of tests to determine the local EPS material properties. The laboratory testing program consisted of unconfined compression (UC) tests to determine stiffness and modulus of elasticity (young s modulus) for EPS by using triaxial machine. As shown in Figure 1a, the direct shear test (DST) to determine the internal shear strength (internal friction angle and cohesion) between EPS Geofoam beads, and modified direct shear test (mdst) to determine the external shear strength (external friction angle) between EPS blocks and other surface. More than 20 modified DST between EPS and other materials (complete testing procedures are found in Abdelsalam et al., 2015). These tests were conducted to characterize the material and interface properties of local EPS and to serve as basis towards the calibration of the hardening soil constitutive model. The reduction factor (Rinter) values for various interface surfaces between EPS and other materials with respect to contact pressure is provided in Figure 1b after AbdelSalam and Azzam (2016). The figure indicates that the relation between Rinter and the contact pressure is non-linear and tends to be constant at high stress levels. This chart is useful for determining the interface parameters required for modeling geotechnical applications that include EPS geofoam. NUMERICAL MODEL A one-meter embedded wall was simulated on the FE software package PLAXIS 2D v.8.6 professional package (Brinkgreve et al., 2006). The finite element analysis for the prototype testing was conducted using the plane strain model. Figure 2a shows the model boundary was totally fixed to surround the entire model. The vertical boundaries were free to move in the vertical direction and fixed in the horizontal direction. The standard soil behind the wall with unit weight around 20 kn/m 3 and the internal friction angle 35 o was replaced with the EPS geofoam with unit weight of around 0.2 kn/m 3 (20 kg/m 3 ), using various EPS thicknesses starting from ration t/h = 0.025, 0.05, 0.09 to 0.13 (where t is the EPS thickness, and h is the wall height). This large range of t/h was used in order to determine the percentage of the reduction in horizontal earth pressure on the embedded wall with respect to EPS thickness. Figure 2b shows the model boundary in the finite element model after the exaction which equals to 45 cm from the top surface of wall. A 15-nodes unstructured deformed mesh was used (see Figure 2c).

5 Figure 1. (a) UC-tests; and (b) Rinter of EPS (after AbdelSalam and Azzam, 2016) Figure 2: FE model with no EPS buffer: (a) geometry; (b) phases; and (c) deformed mesh The constitutive model that used for the EPS, wood and sand was the hardening soil model, linear elastic, and Mohr-Coulomb, respectively. Properties for the EPS were assumed after Abdelsalam et al. (2015), while wood properties were assumed after Robert (2006). The properties and the constitutive models used for all materials in the FE are included in Table 1. Table 1. Material properties used in FE model Material EPS Wood Sand Model Hardening Linear Elastic Mohr-Coulomb Density γ (kn/m 3 ) 0.2 N/A 17 Cohesion C (kpa) 12 N/A 1 Friction angle φ (ᵒ) 33 N/A 31 Initial stiffness Eref (kpa) x x 10 4 Interface factor Rinter 0.56 Rigid Rigid

6 Wall depth (cm) 1 st BUE Annual Conference & Exhibition-BUE ACE1 Cairo, EGYPT 8-10 November 2016 Figure 3a shows the lateral earth pressure where the range of the lateral stress was 6.6 kn/m 2 at the toe of the embedded wall without EPS buffer, but equal to 5.71 kn/m 2 with EPS buffer (t/h=0.025). The lateral stress was reduced with increasing EPS buffer which reach to 2.56 kn/m 2 with EPS buffer (t/h=0.09). Figure 3b shows the lateral deformation of the embedded wall and the lateral deformation in the EPS buffer. a) b) Figure 3. (a) Lateral stress on the embedded wall; and (b) lateral deformation Figure 4 illuminates the effect of EPS buffer on the lateral earth pressure with different wall heights. Wall height is represented on vertical axis lateral earth pressure in horizontal axis. It is observed that the lateral earth pressure increases nonlinear with wall height which is expected. The lateral pressure is around to 4 kn/m 2 by using (t/h= 0.025) of EPS buffer, but it decreases after using (t/h= 0.13) to reach around 2.1 kn/m Lateral Stresses (kn/m 2 ) FEM (t/h=0.025) FEM (t/h=0.05) FEM (t/h=0.09) FEM (t/h=0.13) Figure 4. Lateral earth pressure on embedded wall

7 PHYSICAL PROTOTYPE Prototype Description A physical prototype was vital to verify the results of the PLAXIS 2D model, hence results of the prototype developed by AbdelSalam et al. (in press), assembled in the laboratory of the British University in Egypt (BUE), with the dimensions of 140 cm in length, 60 cm in width, and 100 cm in height. The side walls at the boundaries of the prototype were made of plywood with thickness of 18 mm and were braced along the perimeter using three horizontal struts (5 x 5 cm) and at the corners using four vertical struts. The base of the prototype was made of wood with thickness equal to 5 cm, and directly placed on a concrete slab-on-grade. Figure 5a shows a photo for the prototype boundaries, also shows some details for the front and back boundaries. According to AbdelSalam et al. (in press), the flexible wall of the prototype was made of plywood with thickness equal to 12 mm without bracing, so that the stiffness of the wall is small to mobilize the active earth pressure. As can be seen in Figure 5b, the plywood wall was installed 40 cm from the front boundary to leave 100 cm behind it for the soil to develop the required movement. The wall height was 100 cm, totally fixed from the bottom with the base, and also fixed from both sides with a height of 55 cm. These fixities were made based on manual calculations assuming that the wall imbedded in the soil and has a free length of 45 cm. Hence, the soil backfill was initially placed on both sides of the wall in 5 equal layers of 20 cm, then removed in 10 equal layers (5 cm each) from the passive side to a maximum depth of 45 cm from the wall top. The soil backfill was course sand with same properties summarized in Table m length Corner struts (5 x 5 cm) Course sand (ϕ = 35 o ) Wires from four PS to data acquisition EPS geofoam (t/h = 0.05) Flexible wall made of Plywood (12 mm) 0.6 m width 1.00 m height Plywood (18 mm) Bracing (5 x 5 cm) Displacement gauge (a) Window for soil removal Base (5 cm thickness) (b) 4 Pressure Sensors (PS) Area 3.8 cm x 3.8 cm, thickness 0.46 mm, capacity 980 kpa, and resolution 0.5% Data acquisition soil side PS distribution 25 cm along the plywood wall length from 40 cm the soil side Figure 5. (a) Geometry; and (b) filled with sand (copied from AbdelSalam et al., in press) PS.1 PS.2 PS.3 PS.4 20 cm 10 cm

8 Before testing, four square-shaped force sensing resistors (or pressure sensors, PS) were installed along the wall from the active side. The shape and distribution of the sensors are shown in Figures 5a and 5b. The PS active area was 38.1 x 38.1 mm and thickness was 0.46 mm, which is almost flushed with the wall face. The PS pressure sensitivity range was from 0.1 kg/cm 2 to 10 kg/cm 2 with a resolution of 0.5%, hence can work properly within the expected earth pressure range. The four PS sensors were connected with an in-house data acquisition shown in Figure 5a and recordings were saved to the computer. Moreover, a displacement dial gauge was installed at the top of the flexible wall and located at mid-wall width. The dial gauge was installed to monitor the horizontal displacement for every test configuration. Porotype results Five configurations were tested using the prototype and that for verification purposes required for the 2D model. In the first configuration the lateral pressure on the wall stem was measured with no EPS buffer (control test), whereas in the other four configurations EPS inclusions with t/h equal to around 0.025, 0.05, 0.10, and 0.13 (density = 20 kg/m3, to be similar to the EPS used in the 2D model). Figure 6a shows the relationship between the lateral stresses to the wall depth using 50 mm EPS geofoam buffer (lateral stresses along the horizontal axis and the wall depth along the vertical axis). The figure includes ten curves where each curve represent the lateral stresses for each excavation stage, in addition to the straight line calculated based on Rankine (1857). As can be seen from the figure, the lateral stresses increased by increasing the excavation level until reaching the maximum designed excavation depth, then the curves started to decreasing at the bottom third of the wall height. Therefore, by using the 50 mm thickness of the EPS, the prototype result were very close the theoretical earth pressure by Rankine along the upper two thirds of the wall, then the stress along the bottom portion significantly dropped. By using a larger EPS buffer with thickness equal to 90 mm, a significant reduction in the earth pressure was noticed compared to Rankine. This reduction can be seen in Figures 6b, reached around 25% at depth 65 cm along the wall, then was reduced to around 60% at the wall base. By increasing the EPS thickness to 130 mm (see Figure 6c), the reduction in lateral stress at wall depth 65 cm was 37% and at wall base was 77%. Hence, using the EPS buffer can significantly reduce the earth pressure acting on flexible walls, which agrees with AbdelSalam et al. (in press), and Ertugrul and Trandafir (2013), and is justified because a large compression occurred within the

9 bottom segment of EPS leading to movement within the bottom portion of the soil, which developed shear stresses on the sides of the moved soil portion, and hence part of the lateral stresses was carried by friction with the neighboring soils. (a) EPS 50 mm (b) EPS 90 mm (c) EPS 130 mm Figure 6: Prototype results using EPS with various thicknesses MODEL VERIFICATION Figure 7 shows a comparison between the prototype results and outcomes from the 2D numerical model using EPS thickness of 25 mm, 50 mm, 90 mm and 130 mm. By comparing outcomes from the 2D model with the results from the prototype, it was noticed that there was a small difference of about 10% within the upper two-thirds of the wall, whereas this difference started to increase at the wall base. Apparently the 2D model did not capture the reduction in the stress along the wall base. In general, the curve from the 2D model closely followed the same pressure reduction profiles provided by the prototype. The main observation after increasing the EPS thickness could be the significant reduction in the lateral pressure compared with the no buffer case. Hence, the 2D model showed an acceptable accuracy along the upper two-thirds of the wall but did not captured the rebound in the lateral stresses within the lower third of the wall. Therefore, the reduction in the lateral pressure using outcomes from the 2D model was verified along the upper portions of the wall, and is comparable to outcomes from a physical prototype. CONCLUSIONS In this study, the EPS geofoam buffers was first introduced behind diaphragm walls (flexible wall type) as a new approach in the design that can reduce the applied earth pressure. A summary of the major findings is presented below.

10 (a) EPS 25 mm (b) EPS 50 mm (c) EPS 90 mm (d) EPS 130 mm Figure 7: Comparison between FE and prototype using EPS with various thicknesses 1. Using an EPS inclusion to act as a buffer between flexible diaphragm walls and soil can reduce the acting lateral earth pressure mainly due to soil arching mechanism. 2. Reduction in the lateral pressure can reach up to 37% by using EPS buffer with thickness 130 mm (i.e., thickness to wall height ratio t/h = 0.13). 3. The numerical 2D model provided relatively conservative and accurate results compared with the physical prototype outcomes. However, more research is needed to determine the method of statement and EPS installation technique with diaphragm walls. ACKNOWLEDGEMENTS This study is part of the ongoing research project "development of EPS geofoam to improve infrastructure efficiency", which is funded by the national Science and Technology Development Fund (STDF) project number The authors are grateful to the British University in Egypt (BUE), which hosted the tests in its laboratories. Special thanks to Prof. Hani Amin (BUE) for help and fund provided during the prototype instrumentation.

11 REFERENCES AbdelSalam, S. S., and Azzam, S. A. (2016). Reduction of lateral pressures on retaining walls using geofoam inclusion. Geosynthetics International, accepted for publication 6 Feb AbdelSalam, S. S., Azzam, S. A., and Fakhry, B. M. (in press). Reliability and 3D modeling of flexible walls with EPS inclusion. ASCE, International Journal of Geomechanics, special issue on Advance Characterization and Modeling of Geomaterials and Geosystems. Abdelsalam, S., & Azzam, S. (2015). EPS Geofoam to Reduce Lateral Earth Pressure on Rigid Walls. Int. Conf. on Advances in Structural and Geotechnical Eng., Hurghada, Egypt. Athanasopoulos, A. Z., Lamote, K., Athanasopoulos, G. A. (2012). Use of EPS geofoam compressible inclusions for reducing the earthquake effects on yielding earth retaining structures. Soil dynamics and earthquake engineering, vol. 41, pp Brinkgreve, R. B. J., Kumarswamy, S., and Swolfs, W. M. (2015). Plaxis 3D reference manual anniversary edition version 1. ISBN-13: , Plaxis bv, Delft, Netherlands. Brinkgreve, R. B., Broere, W., and Waterman, D. (2006). Plaxis 2D software manual version 8. Delft University of Technology and Plaxis Inc., Delft, Netherlands. Ertugrul, O. L., and Trandafir, A. C. (2013). Lateral earth pressures on flexible cantilever retaining walls with deformable geofoam inclusion. Eng. geology, vol. 158, pp Horvath, J. S. (1994). Expanded polystyrene (EPS) geofoam: an introduction to material behavior, Geotextiles and geomembranes, vol. 13, no. 4, pp Itasca (2008). FLAC: fast lagrangian analysis of continua, version 6. Itasca Consulting Group, Inc., Minneapolis, MN, USA. Karpurapu, R. and Bathurst, R. J. (1992). Numerical investigation of controlled yielding of soilretaining wall structures. Geotextile and geomembranes, vol. 11, no. 2, pp Lutenegger, A. J. and Ciufetti, M. (2009). Full-scale pilot study to reduce lateral stresses in retaining structures using geofoam. Final report, project No. RSCH Vermont DOT, University of Massachusetts, Amherst, MA, USA. Padade, H. H. and Mandal, J. N. (2014). Expanded polystyrene-based geomaterial with fly ash. International Journal of Geomechanics, Vol. 14, No. 6. Robert, J. (2006). Mechanical Properties of Wood-Based Composite Materials. USA. Zarnani, S. and Bathurst, R. J. (2008). Numerical modelling of EPS seismic buffer shaking table tests. Geotextiles and geomembranes, vol. 26, no. 5, pp