International Society for Soil Mechanics and Geotechnical Engineering

INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of the International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE). The library is available here: https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands of papers published under the Auspices of the ISSMGE and maintained by the Innovation and Development Committee of ISSMGE. Behaviour of rigid retaining wall with relief shelves with cohesive backfill

V. B. Chauhan1 and S. M. Dasaka2

1Research Scholar, Department of Civil Engineering, Indian Institute of Technology Bombay, Mumbai 400076, India; email: [email protected] 2Associate Professor, Department of Civil Engineering, Indian Institute of Technology Bombay, Mumbai 400076, India; email: [email protected]

ABSTRACT Present study attempts to investigate the possible reasons behind the failure of a cantilever retaining wall with relief shelves, which is located in the heart of Hyderabad city, India. The height of the failed retaining wall ranges from 10 to 13.9 m and retains a firm to stiff cohesive backfill, and constructed with 5 relief shelves. After few years of construction, a portion of retaining wall of about 20 m length had collapsed and adjoining 20 m length had severely distressed, immediately after the end of a monsoon. From the preliminary post-failure investigation, it is noted that quality of concrete used in construction was satisfactory, and the construction joints were intact. To get more insight about the causes of failure, numerical analysis of retaining wall with relief shelves is carried out in undrained condition of saturated cohesive backfill, as the wall had failed just after monsoon. From the preliminary analysis, it is noted that, though the lateral thrust on the retaining wall in the presence of relief shelves is reduced up to 18%, use of inappropriate magnitude and distribution of lateral earth pressure in the design calculations might have attributed to the failure of the wall. INTRODUCTION A retaining wall is a structure, which is designed and constructed to resist the lateral pressure of soil, to support vertical or near vertical backfills. There have been situations where high retaining walls are required to resist the lateral earth pressure. Reinforced soil walls may be a possible solution for such cases, but for construction of such walls, a well graded granular material is preferable due to its higher shear resistance and good soil reinforcement interaction, where undrained conditions would prevail. So, availability of a suitable backfill material is a prerequisite for its suitability in reinforced soil wall construction. One alternative to tackle such issues is to reduce the lateral thrust on the wall, which would obviously reduce the sectional dimensions of the wall and cost of the project. A pressure relief shelf is a thin horizontal cantilever platform of finite width, extending into the backfill at right angles, throughout the length of the retaining wall, constructed monolithically with the stem of the retaining wall. Number of such shelves is constructed at regular spacing along the height of the wall. A few researchers previously proposed this technique with limited theoretical studies but without systematic analysis and proper validation, and demonstrated that provision of relief shelves can reduce lateral earth pressure on retaining walls and subsequently increase the stability of the retaining (Jumikis 1964; Chaudhuri et al. 1973; Banerjee 1977 and Bowles 1997). Chaudhuri et al. (1973) demonstrated the benefit of single relief shelf on the reduction of total

lateral thrust on a cantilever wall, through stability analysis of wedges as well as small-scale physical model tests. Through small scale model tests, it was showed that wall with relief shelf can retain larger height of sand just prior to the incipient overturning compared to wall without relief shelf (Chaudhuri et al. 1973). Through a series of model tests on instrumented wall, Yoo et al. (2012) and Moon et al. (2013) showed that distribution of lateral earth pressure on the retaining with relief shelves is a compound function of width and position of relief shelf on wall. Similarly, through the finding of model tests, it is noted that when the relief shelf is located below a certain depth, it could not contribute much to the lateral earth pressure reduction in upper part of wall (Liu et al. 2011). Also some recommendations were laid for optimum ratio of location to width of relief shelf for possible distribution of lateral earth pressure on upper part of wall. Analogous to proposed lateral earth pressure below the relief shelf (Jumikis 1964; Chaudhuri et al. 1973 and Bowles 1997), zero earth pressure is reported just below the relief shelf from the findings of model study of pile-supported cantilever retaining wall with single relief shelf (Liu et al. 2013). To study the effectiveness of various width and location of one/two relief shelves, authors have also investigated lateral earth pressure on the retaining with relief shelves by conducting a physical model test in laboratory for 0.6m high instrumented wall with relief shelves (at different positions with varying width) and noted that provision of relief shelf contributes to the reduction of earth pressure on the wall and make the design economical (Chauhan and Dasaka 2016 and Khan et al. 2016). A case study of failure of a 10-13.9 m high cantilever retaining wall with relief shelves located in Hyderabad, India, had been reported. The above structure had failed after few years of construction, and cracks on the stem of retaining wall just below one of the relief shelves were noted, as shown in Fig. 1. The forensic studies reveal that quality of concrete

Figure 1. Cantilever retaining wall with relief shelves in Hyderabad, India. used in the wall construction was very satisfactory, and construction defects were completely ruled out. To get more insight into the causes of failure, Chauhan et al. (2016) conducted numerical analysis of retaining wall with pressure relief shelves considering cohesionless

backfill. It was reported that larger width of relief shelves, i.e. 2.5 m, used in the above study, might have significantly increased the stresses in the stem of retaining wall just below the relief shelves, leading to unanticipated high tensile and compressive stress on the faces of stem of wall just below one of the relief shelves. These unanticipated stresses might have been neglected in the designs, resulting in failure/distress of retaining wall. As the wall had failed just after monsoon, so poor drainage and earth pressure generated due to saturated backfill may be a probable reason to failure. In order to investigate the possible reason behind the failure of wall, this study is extended with saturated cohesive backfill material in undrained analysis. The present study is aimed at understanding the behaviour of such walls having cohesive backfill and ascertain the effectiveness of relief shelves to reduce lateral thrust and getting proper insight into the associated mechanisms involved in the failure of wall. INVESTIGATION OF FAILURE OF A RIGID RETAINING WALL WITH RELIEF SHELVES Failed retaining wall at Hyderabad, India, with relief shelves has been analysed in FLAC3D. Sectional dimensions of the wall (Fig. 2b) were obtained from the forensic report available with the client (Chauhan et al. 2016). As the soil (backfill and foundation) and wall properties are not available to the authors, so an acceptable range of material properties were taken from Singh and Babu (2010) and Chauhan et al. (2016) respectively, as shown in Table 1.

Figure 2. Cantilever retaining wall with relief shelves, Hyderabad (a) result of numerical analysis (b) sectional dimensions (m). From the numerical analysis, it is found that retaining wall has failed and an attempt is made to capture the progressive failure of wall to understand the reason behind the inception of failure. It is found that due to use of high width of relief shelves, third relief shelf from the backfill surface is severely stressed at the wall and relief shelf junction due to very high bending

stresses at the junction (Fig. 2a), due to which wall portion above to this relief shelf has displaced significantly compared to lower part of wall stems (Fig. 3a). Table 1. Material properties (Chauhan et al. 2016 and Singh and Babu 2010) Property Backfill and Foundation soil Retaining wall Bulk unit weight (kN/m3) 19.0 25.0 Modulus of elasticity (kN/m2) 3 × 104 2.9 × 107 Poisson’s ratio 0.3 0.15 Friction angle (degrees) 27.5 -- Cohesion (kN/m2) 10 --

Also a sign of succeeding displacement started at the relief shelf next below it. Due to this wall movement, a subsequent phase of movement in backfill can be observed in Fig. 3b and 3c. Also a stress reversal of generated stresses on faces of wall (contrary to conventional rigid retaining cantilever walls) is observed near the wall stem junction similar to that observed in the study of same wall with unsaturated cohesionless backfill (Chauhan et al. 2016). It is noteworthy that displacement at wall-shelf junction (Fig. 3a) is similar to failure of wall and crack below one of relief shelf (Fig. 1). These unanticipated stresses might have been ignored during the design of the retaining wall, which resulted cracking of the stem of retaining wall.

Figure 3. Cantilever retaining wall with relief shelves, Hyderabad (a) displacement of wall started (b) progressive displacement in backfill system (c) overall failure of wall system. MODELLING OF RETAINING WALL WITH RELIEF SHELVES To provide a possible solution for the failed retaining wall with relief shelves with cohesive backfill, a cantilever retaining wall having a height of 14.2 m has been chosen for the present study (Fig. 4). Five cantilever relief shelves of same widths are provided at different heights of

the wall (Chauhan et al. 2016). Cohesive soil has been selected as backfill and foundation soil (same as shown in Table 1). Width of relief shelf is varied from 0.6 m to 1.5 m to examine the reduction of lateral earth pressure and total thrust. Length of wall is considered as 1.0 m for analysis. Conventional retaining wall without relief shelves (Fig. 5a) is hereafter referred to as RS 0.0. Retaining wall with relief shelves is shown in Fig. 5b, where B represents width of relief shelf which is varied as 0.6 m, 0.9 m, 1.2 m and 1.5 m, having thickness of 0.3 m and referred to as RS 0.6, RS 0.9, RS 1.2 and RS 1.5.

Figure 5. Sectional dimensions of retaining wall (a) without relief shelf and (b) with relief shelves (Chauhan et al. 2016) and (c) numerical grid of rigid retaining wall with relief shelves (not to scale). The rigid wall is modelled as elastic material and backfill material is modelled as an elasto- plastic material following Mohr-Coulomb failure criterion. Material properties considered in the analysis are shown in Table 1. Fig. 5 shows the numerical grid considered to simulate the rigid retaining wall having static surcharge of 30 kPa. Fixed boundary condition at bottom of foundation and roller boundary condition at vertical end of soils are chosen to represent field conditions. Numerical model described above is validated with the experimental findings of Ertugrul and Trandafir (2011) and discussed in Chauhan et al. (2016). RESULTS AND DISCUSSION In the present analysis, rigid retaining walls with five relief shelves provided at different heights of wall having equal widths are analysed with FLAC3D. The lateral earth pressure distribution, contact pressure below base slab, total lateral thrust and deflection of relief shelves are analysed and discussed below. Contact pressure below base slab Variation of contact pressure below base slab for all retaining walls considered in the present study is shown in Fig. 6. Contact pressure is marginally lower in case of walls with shelves. With

increase in width of relief shelf, contact pressure below the base slab has reduced by maximum 2% only.

Figure 6. Contact pressure below the base for various retaining walls. Lateral earth pressure and total thrust on the retaining wall Distribution of earth pressure on all walls with and without relief shelves have been studied and shown in Fig. 8. Provision of five relief shelves has divided the whole retaining wall into six small segments.

Figure 7. Lateral earth pressure on the wall for rigid retaining wall with relief shelves.

Table 2. Total thrust and reduction in thrust on retaining walls Wall type RS 0.0 RS 0.6 RS 0.9 RS 1.2 RS 1.5 Total thrust (kN/m) 19074 17013 16858 15609 16123 % Reduction in thrust ------10.8 11.6 18.2 15.5

From Fig. 7, it can be observed that lateral earth pressure (total stress analysis) in top first segment has not changed with width of relief shelf which is in line with Liu et al. (2011), which suggests that relief shelf does not much participate in reduction of earth pressure in upper wall

section. In lower portions of wall, earth pressure reduced with increase in width of relief shelf, which can be attributed to the fact that a great portion of overburden and surcharge is carried by uppermost relief shelf. Total thrust from above earth pressure distribution is calculated and shown in Table 2. A noteworthy amount of total thrust reduction is obtained by provision of relief shelves. A range of 11-18% of total thrust reduction is achieved by provision of relief shelves having saturated cohesive backfill with static surcharge of 30kPa.

Lateral displacement of retaining walls and deflection of relief shelves A typical displacement of wall away from backfill has been shown in Fig. 8. It can be seen that provision of relief shelves to the wall has marginally reduced the maximum lateral displacement of the wall from 25.2 mm (wall without relief shelf) to 24.9-24.2 mm (walls with relief shelf).

Figure 8. Contour of lateral displacement of RS 0.9 and summary of maximum displacement of rigid retaining wall with relief shelves With increase in width of relief shelf, maximum displacement of retaining walls has been reduced, which is due to the reduction of total thrust on wall and increased weight of wall due to relief shelves. Maximum deflection of all relief shelves from top to bottom are compared and summarized in Table 3. The notations S1, S2, S3, S4 and S5 represent the relief shelves from top to bottom of retaining wall. Deflection of relief shelves from top to bottom of wall has reduced and found minimum for bottommost relief shelf for all retaining walls with relief shelves. Deflection of relief shelves has significantly increased where the width of relief shelf is greater than 1.2 m. Table 3. Maximum deflection (mm) of relief shelves for various retaining walls Relief Shelf RS 0.6 RS 0.9 RS 1.2 RS 1.5 S1 1.39 2.10 2.93 3.45 S2 1.17 1.94 2.52 3.26 S3 1.02 1.56 2.23 2.84 S4 0.83 1.31 1.85 2.34 S5 0.78 1.20 1.73 2.02

This observation would restrict maximum width of relief shelves to 1.2 m. Larger widths of relief shelves lead to excessive deflection due to its own weight, which may further increase due to creep. Among all the cases of retaining wall with relief shelves, RS 1.2 provides maximum benefit, without leading to excessive deflection of relief shelves. CONCLUSIONS The study involves comprehensive finite difference numerical analysis to examine the possible reason of failure of retaining wall with relief shelves. It is found that use of larger width of relief shelves has significantly increased bending stress in relief shelf as well as on the faces of stem of wall just the relief shelves. This unanticipated stresses might have been neglected in the designs, resulting in failure/distress of retaining wall. From the present study, it is noted that this technique of reducing earth pressure on retaining walls may prove economical. Among all the cases of retaining wall with relief shelves, RS 1.2 proves viable, without leading to excessive deflection of relief shelves. The following conclusions are drawn from the present study. 1. Retaining walls with relief shelves can considerably reduce the total thrust on wall with even cohesive backfill. For the present study under prescribed surcharge, a total reduction is thrust is noted in range of 11-18%. 2. Among all walls considered in the present study, using relief shelves of width 1.2 m will be effective without leading to excessive deflection of relief shelves. 3. Deflection of relief shelf is proportional to the width of relief shelf, and it also decreases from top shelf to bottom shelf for a given retaining wall with relief shelves. REFERENCES Bowles, J.E. (1997). Foundation analysis and design, 5th Edition, McGraw-Hill, Singapore. Chaudhuri, P.R., Garg, A.K., Rao, M.V.B., Sharma, R.N., Satija, P.D. (1973). “Design of retaining wall with relieving shelves”, IRC J. 35(2), 289 - 325. Chauhan, V.B. and Dasaka, S.M. (2016). “Reduction of Lateral Earth Pressure Acting on Non- yielding Retaining Wall using Relief Shelves” Proc. Int. Geot. Engg. Conf. on Sustainability in Geot. Eng. Practices Related Urban Issues, Mumbai, India. Paper ID-34 Chauhan, V.B., Dasaka, S.M., Gade, V.K. (2016). “Investigation of failure of a rigid retaining wall with relief shelves”. Jap. Geot. Society, 10.3208/JGSSP.TC302-02. Ertugrul, O. L. and Trandafir, A.C. (2011). “Reduction of lateral earth forces acting on rigid non- yielding retaining walls by EPS geofoam inclusions”, J. Mater. Civil Eng., 23(12), 1711- 1718. Jumikis, A.R. (1964). Mechanics of soils, D. Van Nostrand Company Inc, Princeton NJ. Khan R., Chauhan V.B. and Dasaka S.M. (2016). “Reduction of lateral earth pressure on retaining wall using relief shelf: A numerical study”, Int. conf. soil env., Bangalore, India, Paper no-117. Liu, G., Hu, R., Pan, X., Liu Y. (2011): Model tests on earth pressure of upper part wall of sheet pile wall with relieving platform Rock and Soil Mechanics, 32(2), 103-110. Liu, G., Hu, R., Pan, X., Liu Y. (2013): Model tests on mechanical behaviors of sheet pile wall with relieving platform. Chinese J. Geotech. Eng., 35(1), 94-99. Singh, V.P. and Babu, G.L.S. (2010). “2D Numerical Simulations of Soil Nail Walls”, Geot. and Geol. Engg., 10.1007/s10706-009-9292-x.