Suitability of Different Materials for Stone Column Construction Dipty Sarin Isaac and Girish M. S. Department of Civil Engineering College of Engineering, Trivandrum, Kerala diptyisaac@yahoo.com, girishmadhavan@yahoo.com ABSTRACT The influence of column material in the performance of stone column is studied through laboratory experiments on model stone columns installed in clay. Five reinforcement materials were studied: stones, gravel, river sand, sea sand and quarry dust. Load versus settlement response was determined. The grain size of the stone column material is one of the main controlling parameters in the design of stone columns. The unreinforced soil under the same loading condition was analysed. It was found that stones are the most effective stone column material. Quarry dust, though a waste product is effective in improving the load deformation characteristics of the soil used. There is no significant difference in the load deformation behaviour of soil with stone columns using river sand and sea sand. Experimental study on behaviour of group of three columns and seven columns was also conducted. A finite element analysis using -noded triangular elements with the software package PLAXIS was also carried out. KEYWORDS: stone column; clay; quarry dust; sand; gravel; stones; load; settlement; model test. INTRODUCTION Among the various methods for improving in situ ground conditions, stone columns are considered one of the most versatile and cost-effective ground improvement techniques. Stone columns have been used extensively in weak deposits to increase the load carrying capacity, reduce settlement of structural foundations and accelerate consolidation settlements due to reduction in flow path lengths. Another major advantage with this technique is the simplicity of its construction method. The type and grain size of stone column material is one of the controlling parameters in the design of stone column. Five materials i.e. stones, gravel, river sand, sea sand and quarry dust, which are stiffer and stronger than the ambient soil were used as column material. The degree of improvement of a soft soil by stone columns is due to two factors. The first one is inclusion of a stiffer column material (such as crushed stones, gravel, etc.) in the soft soil. The second factor is the densification of the surrounding soft soil during the installation of the stone column itself and the subsequent consolidation process occurring in the soft soil before the final loading of improved soil. A detailed experimental investigation was carried out on a single stone column and group of seven and three columns to study the improvement achieved. Spacing of the column also plays an important role in the performance of stone columns.
Vol. [9], Bund. M LITERATURE REVIEW Several researchers have worked on theoretical, experimental and field study on behaviour of stone columns. Ambily and Gandhi () carried out a detailed experimental study on behaviour of single column and group of seven columns by varying parameters like spacing between the columns, shear strength of soft clay and loading condition. Finite element analysis has also been performed using -noded triangular elements with software package PLAXIS. Malarvizhi and Ilamparuthi () studied load versus settlement response of the stone column and reinforced stone column i.e. geogrid-encased stone column in the laboratory. Load tests were performed on soft clay bed stabilized with single stone column and reinforced stone column having various slenderness ratios and using different type of encasing material. Andreou et al. () studied the influence of the main controlling parameters in the design of stone columns through a series of laboratory experiments. The effect of drainage conditions, the grain size of the stone column material, the confining pressure of the soil and the rate of deformation were investigated. Triaxial compression tests were performed on composite soil specimens of soft kaolin clay. The present study aims to determine the most effective material that can be used for stone column construction and to compare the suitability of different materials. River sand is scarce and costly than sea sand. Gravel is also not readily available. Quarry dust is a waste product and is easily available. This study provides guidelines for the selection of proper stone column material which can be used effectively and economically. EXPERIMENTAL STUDIES Materials Six basic materials used for this study are clay, quarry dust, sea sand, river sand, gravel and stones. The clay used was collected from Kuttanad (Champakkulam) in Alappuzha district. In order to maintain uniformity of test results block sample was taken at a depth below m. Particle size distribution is shown in Fig.. The other properties are specific gravity=., liquid limit=.%, plastic limit=.%, maximum dry density=.kn/m, and optimum moisture content=.%. Sea sand was compacted to a dry density of kn/m while constructing stone columns for the experiments. Properties of sea sand used are specific gravity=., maximum dry density=.kn/m, minimum dry density=.kn/m, D=.mm, Ф=9., Cc =.9, and Cu =.. River sand was compacted to a density of. kn/m while constructing stone columns for the experiments. Properties of river sand used are specific gravity=., maximum dry density=.9kn/m, minimum dry density=.kn/m, D=.mm, Ф=9., Cc =., and Cu =.. Quarry dust is a cohesionless material which consists mainly of sand size particle and specific gravity from. to.. Quarry dust was compacted to a density of. kn/m while constructing columns for the experiments.properties of quarry dust are specific gravity=.9, D=.mm, Cc =.9, and Cu =.. Gravel was compacted to a density of. kn/m for constructing stone column.grain size distribution is shown in Fig.. Crushed stones (aggregates) of sizes between and mm have been used to form stone column. The stones were compacted to a density of.kn/m while constructing stone columns for the experiments. Grain size distribution is shown in Fig..
Vol. [9], Bund. M Procedure All experiments were carried out on a mm diameter stone column surrounded by the required soil in a cylindrical tank of mm height and mm diameter to represent the required unit cell area of clay around each column. For group of columns, a tank of mm height and mm diameter was used. Clay was filled in the tank at field water content. Care was taken to ensure that no significant air voids were left out in the test bed. A thin coat of grease was applied along the inner surface of tank wall to reduce friction between clay and tank wall. The centre of the cylindrical tank was properly marked and a PVC pipe of mm diameter was placed at the centre of the tank. Around this pipe clay bed was formed. The clay layer was tamped frequently and gently to expel air during the process of filling. Slight grease was applied on both inner and outer surface of the pipe for easy withdrawal without any disturbance to the surrounding soil. Required stone column material was carefully charged in the tube in three layers to achieve required density. The PVC tube was withdrawn to certain level and charging of stones for the next layer was continued. The operations of charging of stones, compaction and withdrawal of tubes were carried out simultaneously. For installation of group of columns, same procedure was adopted. Enough care was taken to keep the pipes in vertical position. In clay bed the stone columns were prepared from edges towards the centre. Arrangement of columns in seven group and three group column tests are shown in Fig. (a). The tests were conducted on both three column and seven column groups. A typical test arrangement for single column test and group column test are shown in Fig. (b). The stone column was extended to the full depth for a height mm so that l/d ratio (length of column/diameter of the column) is a minimum of. which is required to develop full limiting axial stress on the column. Vertical stress was applied over the entire tank area. The load was applied through a proving ring at a constant displacement rate of.mm/min. A proving ring was used to measure the load and a dial gauge is used to measure the deformation. Load was applied through an mm thick mild steel plate. Percentage finer 9... Particle size(mm)[log scale] Figure : Grain size distribution of clay
Vol. [9], Bund. M Percentage Finer Particle size(mm)[log scale] Figure : Grain size distribution of gravel Percentage finer Particle size(mm)[log scale] Figure : Grain size distribution of stones Analysis of Stone Columns The analysis was carried out using the commercially available finite-element program PLAXIS, to compare the load settlement behavior with the model test. Properties of different materials are shown in Table. An axisymmetric analysis was carried out using Mohr-Coulomb criterion. A drained behaviour was assumed for all materials. Fifteennoded triangular elements were used for meshing. Along the periphery of the tank (interface between the soft clay and the cylindrical surface of the unit cell), radial deformation was restricted but settlement was allowed. Along the bottom of the tank both radial deformation and settlement were restricted. The basic axisymmetric finite-element mesh and boundary conditions used to represent the stone column and the surrounding clay and the typical deformed mesh for single column is shown in Fig. (a). Analysis for a group of seven columns was also carried out as shown in Fig. (b) using an axisymmetric model with surrounding six columns replaced by a ring having equivalent thickness and properties of that material.
Vol. [9], Bund. M Materials E (kn/m ) Table : Properties of Materials Used c µ u φ (kn/m ) (degree) γ dry (kn/m ) γ bulk (kn/m ) Clay.. -.. Quarry dust. -. - Sand. - 9. - Gravel. -. - Stones. -. - TEST RESULTS Typical diagrams of variation of load with displacement were drawn for all the testing conditions. The loaddisplacement curves are, in general, nonlinear. Load Deformation Characteristics of Clay Treated with Single Stone Column Load tests were carried out on Kuttanad clay and load deformation curves were plotted for untreated clay and clay treated with stone column made of five column materials i.e. quarry dust.sea sand, river sand, gravel and stones designated as m, m, m, m and m respectively. These curves are shown in Fig..It is found that stones are most effective as stone column material compared to other materials. Gravel is more effective than sand. River sand perform better than sea sand as the load deformation curve is higher for river sand as the load is increased. Quarry dust, though a waste product is effective in increasing load carrying capacity of clay. Hence it can be economically used as it is cheap and easily available. Figure (a): Arrangement of seven and three columns in a group
Vol. [9], Bund. M Figure (b): Test Setup for Single Column and Group column Test. Figure (a): Finite-element discretization for clay with single column, Typical deformed mesh Figure (b): Finite-element discretization for seven group column
Vol. [9], Bund. M Load(kN)...... untreated m m m m m Figure : Load Settlement Curve for Clay with Single Stone Column Load Deformation Characteristics of Clay Treated with Column Groups Load-deformation curves for seven column group and three column group showed similar trend as shown in Fig. (a), (b), (c) and (d). It is found that load deformation characteristics improved using group of three columns and much using group of seven columns. Tests were conducted for two spacings;.d and d.it is found that load carrying capacity increased as spacing between the columns is decreased. Comparison of Laboratory Tests and FEM Analysis Fig. (a) and Fig. (b) shows typical axial stress versus settlement behaviour for improved and unimproved grounds based on model tests as well as finite-element analysis for single column. Fig. 9(a) and Fig. 9(b) shows stress versus settlement relation for group of seven columns from the two methods(s=d). Comparison is also done with decreased spacing(s=.d) and the graphs are shown in Fig. (a) and Fig. (b). Results from experimental and finite-element analysis matched well regarding stress-settlement relationship. Load(kN) untreated m m m m m Figure (a): Load settlement curve for clay with group of seven columns(s=d)
Vol. [9], Bund. M Load(kN) Untreated m m m m m Figure (b): Load settlement curve for clay with group of seven columns(s=.d) Load(kN)... untreated m m m m m Figure (c): Load settlement curve for clay with group of three columns(s=d) Load(kN) untreated m m m m m Figure (d): Load settlement curve for clay with group of three columns(s=.d)
Vol. [9], Bund. M 9 Stress(kN/ m ) untreated(exp) m(exp) m(exp) untreated(plaxis) m(plaxis) m(plaxis) Figure (a): Comparison of stress settlement relation for clay with single column Stress(kN/m ) m(exp) m(exp) m(exp) m(plaxis) m(plaxis) m(plaxis) Figure (b): Comparison of stress settlement relation for clay with single column Stress(kN/m ) untreated(exp) m(exp) m(exp) untreated(plaxis) m(plaxis) m(plaxis) Figure 9 (a): Comparison of stress settlement relation for clay with group of seven columns(s=d)
Vol. [9], Bund. M Stress(kN/m ) m(exp) m(exp) m(exp) m(plaxis) m(plaxis) m(plaxis) Figure 9 (b): Comparison of stress settlement relation for clay with group of seven columns(s=d) Stress(kN/m ) untreated(exp) m(exp) m(exp) untreated(plaxis) m(plaxis) m(plaxis) Figure (a): Comparison of stress settlement relation for clay with group of seven columns(s=.d) Stress(kN/m ) m(exp) m(exp) m(exp) m(plaxis) m(plaxis) m(plaxis) Figure (b): Comparison of stress settlement relation for clay with group of seven columns(s=.d)
Vol. [9], Bund. M CONCLUSION The following conclusions are drawn from the present study:. Inclusion of stone columns considerably improves the load deformation characteristics of Kuttanad clay.. Among the different stone column materials used, stones are found to be more effective from single column test and group column test.. Quarry dust, though a waste product is effective in improving the load deformation characteristics of the soil used and its performance is comparable with that of sand. Hence quarry dust can be economically and effectively used for stone column construction as it is cheap and easily available.. River sand is more effective than sea sand.. Gravel is more effective than sand in general, though river sand behaves similar to gravel in some cases.. Spacing of the column play an important role in affecting the load deformation characteristics. Effectiveness increases as spacing decreases.. Stress-settlement response is predicted by the finite element method and found matching with experimental results. REFERENCES. Ambily, A.P., and Gandhi, S.R., Behaviour of Stone Columns Based on Experimental and FEM Analysis, Journal of Geotechnical and Geoenvironmental Engineering, vol.,, pp.-.. Andreou, P., Frikha, W., Canou, J., Papadopoulos, V., and Dupla, J.C., Experimental Study on Sand and Gravel columns in Clay, Ground Improvement, vol.,, pp.9-9.. Black, A.J., Sivakumar, M.R., Madhav, M.R., and Hamill, G.A., Reinforced Stone Columns in Weak Deposits: Laboratory Model Study, Journal of Geotechnical and Geoenvironmental Engineering, vol.,, pp.-.. Christoulas, S.T., Giannaros, C.H., and Tsiambao, G., Stabilization of embankment Foundations by using stone columns, Geotechnical and Geological Engineering, vol., 99, pp.-.. Guetif, Z., Bouassida, M., and Debats, J.M., Improved Soft Clay Characteristics due to Stone Column Installation, Computers and Geotechnics, vol.,, pp. -.. Juran, I., and Riccobono, O., Reinforcing Soft Soils with Artificially Cemented Compacted-Sand Columns, Journal of Geotechnical Engineering, vol., 99, pp.-.. Kempfert, H.G., Ground Improvement with Special Emphasis on Column-type Techniques, Int. Workshop on Geotechnics of Soft Soils-Theory and Practice,.. Mitchell, J.K., and Huber, T.R., Performance of Stone Column Foundation, Journal of Geotechnical Engineering, vol., 9, pp.-. 9. Murugesan, S., and Rajagopal, K., Geosynthetic Encased Stone Columns: Numerical Evaluation, Geotextiles and Geomembranes, vol.,, pp.9-.
Vol. [9], Bund. M. Murugesan, S., and Rajagopal, K., Model Test on Geosynthetic-Encased Stone Column, Geosynthetics International, vol.,, pp.-. APPENDIX I. LIST OF NOTATIONS The following symbols are used in this paper C c = Coefficient of curvature C u = Uniformity coefficient D = Effective size φ = Internal friction angle s = Spacing of columns d = Diameter of stone column E = Modulus of elasticity µ = Poisson s ratio γ dr = Dry density γ bulk = Bulk density c u = Undrained shear strength or cohesion 9 ejge