NUMERICAL SIMULATION OF GEOGRID REINFORCED LIGHTWEIGHT GEOMATERIALS ON SOFT GROUND AREA

Transcription

1 Proceeding of the th Asian Regional Conference on Geosynthetics June 17 -, 8 Shanghai, China NUMERICAL SIMULATION OF GEOGRID REINFORCED LIGHTWEIGHT GEOMATERIALS ON SOFT GROUND AREA T. Tanchaisawat 1, D. T. Bergado, Y. P. Lai 3, S. Piyaboon and P. Anujorn ABSTRACT: A full scale test embankment was constructed on soft Bangkok clay using rubber tire chip-sand mixtures as lightweight geomaterials reinforced with geogrid under working stress conditions. The wall/embankment was constructed with solid modular concrete facing units and rock filled gabion boxes on both sloping sides of the embankment. This paper attempts to simulate the behavior of the full scale test embankment using PLAXIS finite element D program. The important considerations for simulation and sensitivity analyses of the behavior of reinforced wall/embankment were the method of applying the embankment loading during the construction process, the assumed soil permeability values during the consolidation process, and the selection of the appropriate models and properties at the interface between the soil and the reinforcement. The effect of settlement predictions are weathered crust thickness and over-consolidation ratio of soft clay because the lightweight backfill loading is much lower compared to conventional backfill. The computed excess pore water pressure is more accurate in shallow depth than deeper depth. The predicted lateral wall movement agreed well with observed field data. Keywords: numerical, geogrid, lightweight geomaterials, Bangkok clay. INTRODUCTION Geosynthetic-reinforced segmental retaining wall or embankment has been well accepted in practice as alternatives to conventional retaining structures; their benefits include sound performance, aesthetics, cost and expediency of construction. This is especially true in soft ground area such as Bangkok, Thailand. Although many geosynthetic-reinforced soil walls have been safely constructed and are still performing well, there are many areas such as alluvial clay or soft clay area that needs in-depth studies in order to better understand the mechanical behavior of this system under more aggressive and harsh environments (Yoo and Song, ). Issues related to the design and factors affecting the performance of reinforced soil have been addressed by many researches in recent times (e.g. Bathurst et al., 5; Park and Tan, 5; Skinner and Rowe, 5; Al Hattamleh and Muhunthan, ; Hufenus et al., ). Also, the behavior of reinforced earth structures has been comprehensively studies through field observation of full scale physical model, laboratory model testing, and numerical simulation (Youwai and Bergdo, ; Bergado and Teerawattanasuk, 7). An alternative method such as a numerical or simulation by means of appropriate methods such as finite-element (FE) or finite-difference (FD) techniques (e.g. Ho and Rowe, 199) is essentially required. Most researches assumed plane strain condition for numerical simulations of reinforced earth structures. Hatami and Bathurst (5) reported a survey of published work on numerical simulation of reinforced soil walls and categorized this work according to: (1) whether numerical models were verified against experimental/field evidence or were simply idealized model; () size of experimental models use for verification of numerical models; (3) quality and extent of measured data reported for each experimental/field case; () assessment of the accuracy of the physical data; (5) simulation of construction sequence and compaction effects; () constitutive models for soil backfill and the availability of laboratory data from which model parameters can be selected; and (7) consideration of load-strain-time effects on mechanical behavior of polymeric reinforcement layers. The focus of this paper is on the numerical modeling using measurements from a series of systematically instrumented full scale test embankment constructed and monitored in plane strain test facility. 1 Lecturer, Department of Civil and Env. Eng g, Kasetsart University CSC Campus, Sakolnakorn, Thailand, tawatchai.t@ku.ac.th Professor, GTE Program, School of Engineering and Technology, AIT, Bangkok, THAILAND, bergado@ait.ac.th 3 Doctoral Candidate, GTE Program, School of Engineering and Technology, AIT, Bangkok, THAILAND, achorsol@gmail.com TenCate Geosynthetics (Thailand) Ltd., Bangkok, THAILAND, somsak@polyfelt.co.th

2 Finite element program Plaxis D Ver. 8. was used to predict the performance of the test embankment during and post construction phases. FULL SCALE TEST EMBANKMENT Subsoil Investigation The test embankment was constructed in the campus of Asian Institute of Technology (AIT). The general soil profile consists of weathered crust layer of heavily overconsolidated reddish brown clay over the top.5 m. This layer is underlain by soft grayish clay down to about 8. in depth. The medium stiff clay with silt seams and fine sand lenses was found at the depth of 8. to 1.5 m depth. Below this layer is the stiff clay layer. Figure 1 summarizes the subsoil profile and relevant parameters. Instrumentation Program The geogrid reinforcement embankment/wall system was extensively instrumented both in the subsoil and within the embankment itself. Since the embankment was founded on a highly compressible and thick layer of soft clay which dictates the behavior of the embankment to a great extent, several field instruments were installed in the subsurface soils. The instrumentation in the subsoil were installed prior to the construction of the geogrid reinforcement wall and consisted of the surface settlement plates, subsurface settlement gauges, temporary bench marks, open standpipe, groundwater table observation wells, inclinometers, dummy open standpipe, dummy surface settlement plates and dummy subsurface settlement gauges (Fig. and 3). Fig. 1 Subsoil profile and relevant parameters Embankment Construction Lightweight geomaterials which made of rubber tire chip-sand mixtures was used as backfill materials. The soil reinforcement was comprised of polyester (PET) geogrid reinforcement material. The facing components were made of segmental concrete block which.3 x.3 x 1. m in dimension. The rubber tire chips were mixed with sand in the ratio of 3:7 by weight. The vertical spacing of the geogrid reinforcement was. m. The backfill was compacted in layers of.15 m thickness to density of about 95% of standard proctor. The sand backfill was used as the surface cover for the rubber tire chips-sand. The thickness of the cover was. m and a non-woven geotextiles was used as the erosion protection on side slope. Hexagonal wire gabions were used on both side of the concrete facing at the front side slopes (Tanchaisawat et al, 7). Fig. Plan view of full scale test embankment with instrumentation Fig. 3 Section view of full scale test embankment with instrumentation NUMERICAL APPROACH The finite element analysis were conducted using a finite element software, PLAXIS D version 8. developed by PLAXIS B. V. (). The program

3 allows a realistic simulation of construction sequences, and the inclusion of reinforcement and interface elements at any stage of the analysis without any significant changes in the input data and finite element mesh. The FE model of reinforced embankment consisted of geogrid reinforcement, soil-toreinforcement interaction and concrete facing elements and their connections. The six-node triangular element was used in model simulation (Fig. ). The soft soil model (SSM), which is similar to the cam clay model, was used to model the behavior of a soft clay foundation. The linear elastic material model was used to model as medium clay layer, concrete block facing, and geogrid reinforcement. The elastic perfectly plastic with Mohr-Coulomb failure criteria was used to model the behavior of weathered crust and backfill material. Fig. Finite element model of full scale test embankment Model Parameters The backfill soil material used in the embankment is silty sand and rubber tire chip-sand mixtures. The elastic, perfectly plastic Mohr-Coulomb model can be used to represent the backfill material. At the construction site, the upper layer 1.5 to.5 m depth consists of the weathered crust, which is heavily overconsolidated. Elastic, perfectly plastic model with constant value of Poisson s ratio has been used for this soil. The Cam Clay model has been widely used for representing stress-strain relationship of the soft Bangkok clay, which is normally consolidated and lightly overconsolidated. In PLAXIS software, Cam Clay type model which is modified Cam Clay Model and Soft Soil model are available. The linear elastic model is used for predicting the behavior of the medium stiff clay layer. The geogrid reinforcements are flexible materials capable only of resisting tensile stresses, and in PLAXIS software this type of material is modeled as geotextile. The elastic perfectly-plastic model was used to simulate the constitutive relations of soil-geogrid interface. The segmental precast concrete was used as embankment facing the size of block is.3 x.3 x 1. m. In this study, the precast concrete facing was modeled as elastic beam. The input data for FEM simulations are tabulated in Table 1. The in-situ stress was introduced in the foundation soil by adopting K o procedure. Then, the backfill was constructed into 1 layers, as was done in the field. The compacted backfill was included in each layer, and the reinforcement was placed on a layer before the next layer was installed. The compacted backfill in a given layer was assigned with the material parameters according to the stress state induced after installing the layer. RESULTS AND DISCUSSIONS Surface Settlements The observed and predicted surface settlements of the test embankment are plotted together in Fig. 7. As expected the predictions from Asaoka (1978) closely followed the observed data while the predictions from one-dimensional method overpredicted. The predicted surface settlements from FEM analysis agreed well with observed data. The computed surface settlement at the beginning of construction is greater than that measured values due to the partially drained behavior effect on the soft clay foundation at the early stages of the construction which can be related to the method of applying the embankment loading on the construction process or stress level. However, the computed settlements after 9 days agreed well with those measured values in the field when the drained behavior is consistent with the actual permeability values. Embankmentheight(m) Time (days) Settlement (mm) Fig. 7 Observed and predicted surface settlements Excess Pore Water Pressure Observed Data Terzaghi 1-D Method Asaoka's Method FEM D Method The excess pore water pressure below the lightweight embankment was obtained from open stand pipe piezometer. Figure 8 shows the measured and

4 Proceeding of the th Asian Regional Conference on Geosynthetics June 17 -, 8 Shanghai, China Table 1 Selected Parameters for Finite Element Modeling Material Depth Model Type γ unsat γ sat k x k y λ * κ OCR c ' φ' E υ ' m kn/m 3 kn/m 3 m/d m/d kn/m deg kn/m Foundation Soil Weathered Crust. -.5 EP(MC) U E-3 1.E Soft Clay SSM U E-.E Soft Clay SSM U E- 5.E Soft Clay SSM U E-.E Midium Clay Elastic U E-3.E Backfill Soil Sand - EP(MC) U E+ 8.E Tire Chip-Sand Mixtures - EP(MC) U E+ 8.E (G).33 Reinforcement Geogrid GX1/3 - Elastic - EA = 5 kn/m (Tension Mode) Interaction Parameter =.95 Facing Modular Concrete Block - Elastic - EA = 1.5 E+1 kn/m EI = 8.75E+8 knm /m w = 7. kn/m/m.15 Note : EP(MC) Elastic Perfectly Plastic (Mohr-Coulomb) EA Axial Stiffness SSM Soft Soil Model EI Bending Stiffness Elastic Linear Elastic w Unit Weight (concrete) U Undrained Condition (G) Shear Modulus after construction at 3 m depth. The maximum pore water pressure of 57 kn/m 3 occurred at 15 days after end of embankment construction. The trend of excess pore water pressure dissipation is an indication of consolidation of soft foundation subsoil in the overconsolidation range when the load is below the maximum past pressure. After 5 days, the excess pore water pressure dissipated very fast with time. The excess pore water pressure decreased to 18 kn/m and 5 kn/m at 3 m and m depths, respectively. The excess pore water pressure becomes constant with time after 1 days from the end of construction. The 1-D method overpredicted the excess pore water pressures than the predictions from the Skempton and Bjerrum (1957). The computed excess pore water pressures from FEM D method reasonably agree well with observed data Time (days) Embankment height (m) Excess pore water pressure (kn/m ) Observed Data One-Dimentional Method Skempton and Bjerrum (1957) Method FEM D Method Fig. 8 Observed and predicted excess pore water pressures at 3 m depth for end of construction and after 13 months of construction. The computed lateral displacements of the wall face generally agreed with the observed field data. In the weakest zone of soft clay at depths of 3 to m below the ground surface, the computed subsoil lateral displacements overpredicted the observed data. Since the lateral wall movement coincides with the results of vertical settlements, the partially drained consolidation process at very early stages of construction is not modeled well by the undrained finite element analysis (Schaefer and Duncan, 1988; Chai and Bergado, 1993) Lateral wall movement (mm) - Observed Data (end of construction) FEM D Method (end of construction) Observed Data (1 years after constructin)) FEM D Method (1 year after construction) Fig. 9 Observed and predicted lateral wall movements CONCLUSIONS Embankment Backfill Weathered Crust Soft to Very Soft Clay Medium Clay Stiff Clay Lateral Wall Movements The comparison between the FEM results and observed field lateral wall movement is shown in Fig. 9 The numerical simulation based on finite element analyses under plane strain condition using PLAXIS computer software were carried out to study the behavior of a lightweight embankment reinforced with

5 geogrid on soft ground foundation. The numerical simulation techniques adopted in this paper captured the overall behavior of the reinforced soil wall/embankment system on a soft foundation through good agreement between the field observations and the simulated values. The important simulation considerations in the FEM analysis consisted of the method of applying the embankment loading during construction process, the selection of an appropriate soil and reinforcement models, the estimation of soil permeabilities soft foundation, and the selection of appropriate parameters at the interface between the lightweight tire chips-sand backfill soil and geogrid reinforcement corresponding to the interaction mechanism. The predicted results were shown to be generally in good agreement with measured settlements, excess pore water pressures and lateral wall movements. ACKNOWLEDGEMENTS Sincere thanks and appreciation go to Assoc. Prof. Dr. Panich Voottipruex from King Mongkut`s Institute of Technology North Bangkok for his kind assistance and usefull discussion. This work was derived mainly from a research project funded by the Royal Thai Government conducted at the Asian Institute of Technology. The facilities provided by the research project are gratefully acknowledged. REFERENCES Al Hattamleh, O., and Muhunthan, B. (). Numerical procedures for deformation calculations in the reinforced soil walls. Geotextiles and Geomembranes, Vol., No. 1, pp Asaoka, A. (1978). Observation procedure of settlement prediction. Soils and Foundations, Vol. 18, No., pp Bathurst, R. J., Allen, T. M. and Walters, D. L. (5). Reinforcement loads in geosynthetic walls and the case for a new working stress design method (Mercer Lecture). Geotextiles and Geomembranes, Vol. 3, No., pp Bergado, D. T. and Teerawattanasuk, C. (7). D and 3D numerical simuations of reinforced embankments on soft ground. Geotextiles and Geomembranes, (in press) Chai, J. C. and Bergado, D. T. (1993). Some techniques for FE analysis of embankment on soft ground. Canadian Geotechnical Journal, Vol. 3, pp Hatami, K. and Bathurst, R. J. (5). Development and verification of a numerical model for the analysis of geosynthetic-reinforced soil segmental walls under working stress conditions. Canadian Geotechnical Journal, Vol., pp Ho, S. K. and Rowe, R. K. (199). Prediction behavior of two centrifugal model soil walls. Journal of Geotechnical Engineering ASCE, Vol. 1, No. 1, pp Hufenus, R., Rueegger, R., Banjac, R., Mayor, P., Springman, S. M. and Bronnimann, R. (). Fullscale field tests on geosynthetic reinforced unpaved roads on soft subgrade. Geotextiles and Geomembrane, Vol., No. 1, pp Park, T. and Tan, S. A. (5). Enhanced performance of reinforced soil walls by the inclusion of short fiber. Geotextiles and Geomembranes, Vol. 3, No., pp PLAXIS B. V. (). PLAXIS D Version 8., Finite Element Code for Soil and Rock Analysis, A. A. Balkema, Delft, Netherlands. Schaefer, U. R. and Duncan, J. M. (1988). Finite element analysis of the St. Alban test embankments. ASCE Geotechnical Special Publication No. 18, pp Skempton, A. W. and Bjerrum, L. (1957). A contribution to the settlement analysis of foundations on clay. Geotechnique, Vol. 7, pp Skinner, G. D. and Rowe, R. K. (5). A novel approach to estimating the bearing capacity stability of geosynthetic reinforced retaining walls constructed on yielding foundations. Canadian Geotechnical Journal, Vol., pp Tanchaisawat T., Bergado D. T. and Kanjananak T. (7). Lightweight geomaterials for bridge approach utilization, Proceedings 1 th Southeast Asian Geotechnical Conference 7, 8-11 May 7, Subang Jaya, Malaysia. Yoo, S. and Song, A. R. (). Effect of foundation yielding on performance of two-tier geosyntheticreinforced segmental retaining walls: a numerical investigation. Geosynthetics International, Vol. 13, No. 5, pp Youwai, S. and Bergado D. T. (). Numerical analysis of reinforced wall using rubber tire chipssand mixtures as backfill material. Computers and Geotechnics, Vol. 31, No., pp