IGC 2009, Guntur, INDIA

A Common Collapse Test for Granular Materials

A COMMON COLLAPSE TEST FOR GRANULAR MATERIALS

A. Trivedi Dean, Faculty of Technology, University of Delhi, and Professor, Delhi College of Engineering, Bawana Road, Delhi110 042, India. E-mail: prof.trivedi@ yahoo.com R. Pathak Assistant Professor, Department of Civil Engineering, Thapar University, Patiala147 004, India. E-mail: rpathak@thapar.edu R. Gupta Lecturer, Department of Civil Engineering, Thapar University, Patiala1470 04, India. E-mail: rgupta@thapar.edu ABSTRACT: A fine grained granular material shows collapse of its pre-wetting structural arrangement upon wetting. Conventionally, the collapse potential of granular material is obtained by oedometer tests. The oedometer is sensitive to the collective effect of varied placement parameters responsible for the collapse of the granular materials but the effect of pre compression and confinement is not well expressed in its results. Therefore, we propose a common collapse test namely Model Plane Strain Collapse Test (MPSCT) where we varied a wide range of placement parameters to obtain collapse of selected granular materials. The experiment shows that the MPSCT closely follows the behaviour of granular fills relative to the oedometer tests. The paper explains the technique of test for MPSCT at the controlled densities. The MPSCT provides an opportunity to allow incidence of collapse on a rising water level. This paper describes a plain strain collapse test for fine granular materials. We describe the results of MPSCT to explain the effect of grain sizes upon collapse of pre-wetting structural arrangements upon wetting. 1. INTRODUCTION 1.1 The Collapse of Granular Materials The granular materials are often found in varied forms in the nature. It can be the fine galactic dust in the interstellar space or a common salt found in the house hold kitchens. One of the most abundant examples of its occurrence on earth is fine grained soils and fine flow-able powders in varied industrial uses. When they come in contact with each-other they organize them self in a variety of structural arrangements. These structural arrangements at varied volume, temperature and pressure conditions show remarkable response on loading. The loose granular materials with specific structural arrangements are vulnerable to volume change on wetting which is associated with the collapse of its pre-wetting structure. The implications of such a collapse can be severe for human habitats. Therefore, an investigation was carried out to examine the factors that capture the collapse of loose ? ll on inundation. The general characteristics of collapsing soils are a sudden and a large volume decrease at a constant stress when inundated with water. According to Lutenegger & Saber (1988) the collapse is associated with the meta-stable structure of a large open and porous fabric of the material. Morland & Hastings (1973) evaluated collapse of the dry porous volcanic tuff. Holtz (1948) suggested that earthen structures such as embankments, road 47 ? lls and structural ?lls may collapse when the placement moisture content is dry of optimum. It has been observed that in a wide range of placement parameters the granular materials remains vulnerable to the collapse on submergence in working stress range. Several studies have indicated that compaction control of many granular fills in the ?eld by usual methods is often poor. It adds to the vulnerability of granular ?lls to a wetting induced collapse. The soils that exhibit collapse have an open type of structure with a high void ratio. According to Barden et al. (1969) the collapse mechanism is controlled by three factors; namely a potentially unstable structure, such as a ?occulent type associated with soils compacted dry of optimum or with loess soils; (2) a high applied pressure which further increases the instability; and (3) a high suction which provides the structure with only temporary strength which dissipates on wetting. Jennings & Knight (1975) indicated that collapse behaviour is also dependent on the clay fraction. Foss (1973) and Reznik (1993) de?ned the magnitude of collapse by, Cp = dh/h0 (1) where Cp is collapse potential; dh is decrease in the height of specimen subsequent to inundation; and h0 is height of specimen before saturation. As per Booth (1975) the collapse settlement exceeding 1% is likely to occur when the initial dry unit weight is less than both 15.7 kN m3 and 85% of that obtained in the modi?ed compaction test. The collapse

A Common Collapse Test for Granular Materials

potential is quanti? able in terms of the volume change that occurs when a soil is submerged with water. It is usually obtained by conducting oedometer tests on soil specimen. The collapse potential is expressed as a change in void ratio on wetting compared to the pre wetting volume of the soil at any stress level, Cp = de/(1 + ei) (2) where de is change in void ratio upon wetting; and ei is void ratio at the beginning of saturation. Knight (1963) proposed the collapse potential at a stress level of 200 kPa selecting ei as void ratio at the beginning of compression. If the Cp is less than 0.01, normally there is no risk of a collapse in the ?eld. Lutenegger & Saber (1988) recommended the use of ei as void ratio before saturation at an applied stress of 300 kPa. Accordingly a soil is slightly collapsible if the value of Cp is below a limit of 0.02. In this study several oedometer tests were performed and the results were veri?ed by specially fabricated model test. A work program was devised to study the in?uence of characteristics, density, stress level, and overconsolidation on the collapse of selected granular materials. The work was planned to examine the in?uence of density and rising water table on the incidence of collapse. Based upon the oedometer and the model tests a new limit of collapse potential was assigned. The behaviour of the selected material was found to be highly correlated with its grain size characteristics (Trivedi & Sud 2002, 2004, 2005 and 2007). They suggested criteria for collapse incorporating model plate load test results along with oedometer results. 2. MODEL PLATE LOAD COLLAPSE TEST The collapse potential obtained by the oedometer is qualitative in nature. The oedometer test takes in to account the one dimensional volume change of the sample strictly con? ned within a ring and two rigid plates. In the ?eld there is a possibility of a vertical deformation accompanied with a volume change, side expulsion, slip, and migration of the ? nes under a slow rising head of water. The collapse of the granular structure below foundations may also depend upon the proximity of the applied loads to the ultimate loads required for a shear failure of the assembly. While one dimensional volume change is quanti?ed by the oedometer, the actual ?eld conditions of side expulsion, the slip displacement of ?nes and the proximity of loads to the shear failure remain far from incorporated. The model plate test simulates the condition of ?eld collapse closely. Reznik (1993) suggested verifying the oedometer results by the plate load test, where an actual wetting is employed. Keeping this point in view, the model plate load collapse tests were conducted in a specially fabricated tank under a plane strain condition. Three tests were performed in varying density states, identi? ed by the oedometer test as non-collapsible and collapsible according to the existing criteria for soils. 48

2.1 Materials and Methods We conducted a set of experiments on granular materials of varying specific gravity, unit weights, and grain sizes in oedometer and MPSCT as given in Table 1. Table 1: Summary of Experimental Program Material type Fine Ash Coarse Ash Silty-Sand 2.2 Model Test Setup A tank of inner plan dimension of 155 by 600 mm and a height of 500 mm was made up of a 10 mm thick mild steel sheet (Fig. 1). The horizontal and the vertical stiffeners were provided to restrict the lateral movement of the vertical plates on the load application. On the front vertical face, a cut slot 100 mm wide was left for the placement of a plate with ? ne perforations. The perforated plates were provided to allow the seepage of water into the deposit from the downward to the upward direction. The perforations were plugged with the ? lter to disallow migration of ? nes outside the inner tank containing saturated granular materials. The displacement of longitudinal stiffeners was monitored by mounting the dial gages on both sides of the tank. The granular materials used in this study is a predominantly silt sized cohesionless material. The rainfall or vibratory densi?cation technique is generally used for the deposition of cohesionless (Da = 1 mm) material on a uniform density. Specific gravity 1.62.0 1.61.7 2.62.7 Test Oedometer Strip Footing Oedometer, Strip Footing Oedometer

Loading Frame Vertical Stiffeners Compacted Fill

Hydraulic Jack Loading Plate

Settlement Gauges Perforated Slots Horizontal Stiffeners Pedestals

To Water Tap

Fig. 1: A Line Sketch of MPSCT (Trivedi & Sud 2004) It was found to be unsuitable for the deposition of the dry granular materials with significant magnitude of fines, where with a little blow of air, granular powder became air borne. A special technique was adopted for the densi? cation in tank by ?rst pouring granular materials in a loose state up to a speci?ed depth by a slow pouring technique. A 55 mm thick

A Common Collapse Test for Granular Materials

layer of a fine granular material was uniformly pressed by a stiff plate to achieve a desired density in the tank. After ?lling the granular materials to a desired density, load test was conducted using a plate 155 mm long, 100 mm wide. The collapse tests were conducted by keeping the model granular materials ?lled tank in a submergence tank 700 by 500 mm plan and 600 mm deep. After the deposition of the granular materials at a desired unit weight, the load was applied through a hydraulic jack, corresponding to a speci?ed stress level and the plate settlement was recorded. The water was allowed into the external tank for the saturation of the granular materials.
100 90 80 70

3. A CRITERIA FOR COLLAPSE It is recognized that a granular material follows a closer packing under a favorable condition of pressure and moisture. This tendency may be quantified in terms of the contact separation parameter (D50/Da) defined in the Figure 3(a) and distance of placement void ratio to the minimum void ratio. It is posed that minimum void ratio occurs in proctor compaction. Therefore a collapsibility factor (F) is defined as, F = (ei emin)/emin (3) where ei is placement void ratio of granular materials and emin is void ratio corresponding to maximum dry density in proctor compaction. Larger the value of F, granular materials is more predisposed to collapse.

Finer by Weight

t h 60 g i e W y 50 b r e n i F 40 % 30 20 10 0 0.001

Range of Mean Sizes Under Investigation

Fig. 3(a): Definition of Contact Separation Parameter D50/Da Figure 3(b) shows a reduction in collapsibility factor F, with mean size in the loosest and the compacted states. Compared to the loosest state, all the granular materials reach almost a common collapsibility level in a compacted state. At 95% degree of compaction a low collapsibility level is reached, which is associated with small volume change on collapse that does reflect insignificant collapse.
3.00
F in loosest state F at Dc=80% F at Dc=90% Anticipated F

0.01

0.1 Grain Size (mm) Gain Size (mm)

10

Fig. 2(a): Grain Sizes Selected for Investigation The settlement of the loading plate was continuously monitored and the load level was constantly maintained. The actual wetting was induced at varied stress levels. The dry settlement up to the collapse load, the collapse settlement on inundation, and the post-collapse wet settlement of granular materials pressed at various densities were noted for all the tests. The post collapse degree of saturation was recorded on the completion of each of the test. Figures 2 (a) and (b) shows a typical grain sizes and load-collapse-settlement plot in MPSCT respectively.
Pressure (kPa) Pressure (kPa) 0 0 1 2 3 20 40 60 80 100 120 140

2.50

Collapsibility Factor

F r o 2.00 t c a F y t i l i b i 1.50 s p a l l o C

Antipated F

F(in loosest state ) = 0.2 (D50/Da)-0.5

Observed F at slow pouring ( loose state)

F (Dc=80%) = 0.4 (D50/D a)-0.1 F (Dc=90%) = 0.2 (D50/Da)-0.1

1.00

0.50

0.00 0.001

0.010
D50/D/Da D50 a

0.100

1.000

Settlement (mm)

) m m ( t n e m e l t t e S

4 5 6 7 8 9 10

Loose Medium Dense Dense

On inundation

Fig. 3(b): Criteria of Collapse Based on Collapsibility Factor (Trivedi & Sud 2004) Moreover, it has practical problem of precise measurement of the volume change. Thus, the variations in the measured collapse at more than 98% degree of compaction may forbid interpretation of any trend. As a result of the above observations the classification of granular materials at lower degree of overburden was found to be appropriate for the evaluation of collapse. The mean 49

Fig. 2(b): A Typical Collapse Settlement in MPSCT (Trivedi & Sud 2004)

A Common Collapse Test for Granular Materials

particle size was seen to control the collapse of granular materials. If the mean size was greater than 1 mm the granular materials were non-collapsible and others were collapsible under specific conditions. The value of collapse potential in the critical range of stress and moisture was 3 to 6 times that of the corresponding dry condition. It suggested susceptibility of a non-collapsible dry granular material to the collapse in partly saturated condition. In order to obtain the value of collapse potential of partly wet granular materials, a multiplier may be applied. The collapsible granular materials were further divided into the granular materials of low, medium and high collapsibility on the basis of their collapse potential. The collapsible and the non-collapsible granular materials were identified using the model plate load collapse test on selected samples. Normally, the weight of a particle of a natural soil of similar grain size is 1.5 to 1.3 times that of ash materials. These soils remain stable at or less than 1% volume change (Cp = 0.01). Being light in weight, the ash material has a propensity to be unstable in the presence of buoyancy which plays a role in the model and the field conditions. Therefore, among the light weight granular materials particles, 0.75% volume change (Cp = 0.0075) triggered collapse failure in the field. Incidentally, 1% volume change of soils is 1.3 times that of the limit recognized for the collapsible ash materials. It was observed that the sand and the coarse ash had a close value of the median size. It being a granular material collected dry; having around 25% particles in silt range, had a higher collapse potential than the sand. It was recognized that all the collapsible granular materials had relatively more fines. Among coarse grained granular materials a scatter in collapse potential was observed. A relationship between collapse potential at 200 kPa (loosely-compacted) and mean particle size is obtained. The collapse potential in relation to contact separation parameter is expressed by, Cp = n. (D50/Da)m (4) where Cp is collapse potential of a granular material, D50 is mean particle size in mm, and Da is reference size, m and n are fitting constants for the granular materials. The fitting curve of observed collapse potential with the mean size changes with specific gravity (Fig. 3c). The equation (4) can predict the value of collapse potential of ashes and silty granular materials. It furnishes a guide by which one may compare the potential severity of the collapse situation according to the grain size. It can be utilized to tell when to expect a problem. The meticulous use of Figure 3(c) in judging just how much collapse will take place in actual field conditions is a matter of further interest. 50

0.07

0.06

0.05

Projected Upper Bound Plot

Collapse Potential

l a i t n e t o P e s p a l l o C

0.04

Trivedi & Sud 2004 MPSCT

0.03

0.02

0.01
Projected LowerBound Plot

Fine

0 0.01 D 50/Da

0.1

D50/Da

Fig. 3(c): Range of Collapse Potential for Granular Materials 4. CONCLUSIONS The collapsibility of granular materials is one of the most important parameters for using granular materials as a load bearing fill material. The present work provides a framework for the assessment of collapsibility of the granular materials. Several oedometer collapse tests have been used to test the collapsibility of granular materials and the results were verified using the MPSCT. The following conclusions are drawn: (a) The collapse potential obtained by the oedometer test is a dependent parameter of several factors such as grain size characteristics, specific gravity, stress level, testing technique, degree of compaction, over consolidation ratio, etc. (b) The observed collapse potential was proportional to the collapsibility factor identified from maximum and minimum void state of the granular materials. The granular materials with more than 50% of the particles in silt size range were found to be collapsible. (c) The generally recognized lower limit of collapse potential for the collapsible soils in the oedometer is 0.01. It was recognized that the limit is function of specific gravity and grain sizes together. A collapse potential of 0.0075 at low density in oedometer, collapsed in model tests at higher densities. Further increase in the density of this granular material arrested the collapse in the model test. The ash based granular materials of mean sizes near reference sizes (1 mm) had a lower collapse potential in oedometer (even at a low density). They did not collapse at all the densities examined in the model test. Therefore, a new paradigm for collapse of the granular materials is recognized as per Figure 3(b & c). ACKNOWLEDGEMENTS The authors thankfully appreciate Prof. V.K. Sud and Prof. Sundar Singh of the erstwhile geotechnical research group of

A Common Collapse Test for Granular Materials

Thapar Institute of Engineering & Technology (now Thapar University) Patiala for their thoughtful inputs in the design of MPSCT. The first author is also thankful to Delhi College of Engineering, Faculty of Technology, University of Delhi, for providing ample space for his research studies. REFERENCES Barden, L., Madedor, A.O. and Sides, G.R. (1969). Volume Change Characteristics of Compacted Clays, JSMFD, ASCE, 96, No. SM4, 3339. Booth, A.R. (1975). The Factors Influencing Collapse Settlement in Compacted Soils, Proceedings, 6th Regional Conf. of Africa on SMFE, 5764. Foss, I. (1973). Red Soils from Kenya as a Foundation Material, Proceedings, 8th ICSMFE, Moscow, Vol. 2.3, 7380. Holtz, W.G. (1948). The Determination of Limits for Control of Placement Moisture in High Rolled Earth Dams, ASTM, Proceedings, 48, 12401248. Jennings, J.E. and Knight, K. (1975). A Guide to Construction on or with Material Exhibiting Additional Settlement Due

to Collapse of Grain Structure, 6th Regional Conf. of Africa on SMFE, 99105. Knight, K. (1963). The Origin and Occurrence of Collapsing Soils, 3rd Regional Conf. of Africa on SMFE, 127130. Lutenegger, A.J. and Saber, R.T. (1988). Determination of Collapse Potential of Soils, GTJ, Vol. 11, No. 3, 173178. Morland, L.W. and Hastings, C.R. (1973). A Void-Collapse Model for Dry Porous Tuff , Engg. Geology, Vol. 7, 8187. Reznik, Y.M. (1993). Plate Load Tests on Collapsible Soils, JGE, ASCE, Vol. 119, No. 3, 608615. Trivedi, A. and Sud, V.K. (2002). Grain Characteristics & Engineering Properties of Coal Ash, Granular Matter, (Springer, Heidelberg), Vol. 4, No. 3, 93101. Trivedi, A. and Sud, V.K. (2004). Collapse Behavior of Coal Ash, JGGE, ASCE, Vol. 130, No. 4, 403415. Trivedi, A. and Sud, V.K. (2005). Ultimate Bearing Capacity of Footings on Coal Ash, Granular Matter, (Springer, Heidelberg), Vol. 7, No. 4, 203212. Trivedi, A. and Sud, V.K. (2007). Settlement of Compacted Ash Fills, Geotechnical and Geological Engineering, (Springer, Heidelberg), Vol. 25, No. 2, 16317.

51