PROBABILISTIC ANALYSIS OF RANDOMLY DISTRIBUTED FIBRE-REINFORCED SOIL

Abstract

Fibre reinforced soil termed as "plysoH" is a composite material obtained by mixing randomly distributed discrete fibres to soil (RDFS). The technique of fibre inclusion in soil, is in general, similar to soil stabilization. The main advantages of RDFS are the simplicity in mixing, maintenance of strength isotropy and absence of potential planes of weakness which may develop parallel to the oriented reinforcement. RDFS can be advantageously utilised as a soil improvement technique with respect to embankment, subgrade/sub-base and in similar other problems. Most of the experimental investigations reported in the literature have been carried out on sands reinforced with discrete fibres. A very limited effort has been made to study the behaviour of fibre reinforced silty sand, silt and clayey soil. Stress-strain modelling of RDFS is complex. Asimple force equilibrium model has been proposed by Gray and Ohashi (1983) for oriented fibres, however the same can not be used directly for randomly distributed fibres. Maher and Gray (1990) proposed a force-equilibrium model with statistical analysis for randomly distributed discrete fibre reinforced sand. The model predicts the orientation and the number of fibres at any plane using the statistical approach for composite materials. The model proposed by Maher and Gray (1990) predicts reasonably well the increase in strength of randomly distributed fibre reinforced sand. However, the width of the shear zone which significantly affects the increase in strength has not been determined for reinforced sand. Also, it is difficult to determine experimentally the orientation and the number of fibres at any plane. The present study is aimed to develop a statistical model to bring out the influence of soil-fibre parameters and confining stress on the strength of RDFS. An exhaustive and well planned triaxial compression test programme has been carried out to study the behaviour I / of RDFS. The influence of fibre properties i.e. weight fraction, aspect ratio, skin friction and modulus of elasticity, soil parameters i.e. particle size, degree of compaction and degree of saturation and confining stress has been studied. Acomputer controlled "GDS Triaxial Testing System" was used to achieve better control and accuracy in the test results. Five different types of cohesionless soils ranging from medium sand to non-plastic silt, and three different types of fibres i.e. polypropylene, coir and bhabar fibres have been used in the investigation. In addition to triaxial tests, California bearing ratio tests were carried out on sand reinforced with both polypropylene and coir fibres, to assess the suitability offibre reinforced sand as subgrade/sub-base material for flexible pavement. Analysis of triaxial compression test results reveal that deviator stress-axial strain curves of fibre reinforced soils show an increasing trend even upto 20% axial strain, indicating a ductile behaviour of the composite material. The stiffness (i.e. secant modulus) of soils is observed to increase at all levels of axial strain due to addition of discrete fibres to soils. The failure envelopes of fibre reinforced soils are observed to be curvilinear having transition at certain confining stress, termed as "critical confining stress", <rcrlt. Thecritical confining stress is noted to decrease with increase in fibre aspect ratio. However, fibre weight fraction and average soil grain size, Dso have no tangible effect on the amount of critical confining stress. Shear strength of fibre reinforced soils increases with increase in fibre content and aspect ratio. For example, increase in strength in terms of major principal stress at failure was of the order of 34%, 65%, 130% and 155% of the strength of unreinforced fine sand, with polypropylene fibres (having l/d = 100) of 0.5%, 1%, 2% and 3% (by weight) respectively, under confining stress of 300 kPa. The percent increase in strength is approximately of the same order for loose and dense conditions of sand reinforced with discrete fibres. California Bearing Ratio (CBR) values of soaked and unsoaked specimens show a marked improvement due to inclusion of fibres, both polypropylene and coir fibres. The CBR values of sand are observed to increase by 50%, 87.5%, 106.2% and 118.7% under unsoaked condition, and 57%, 85.7%, 100% and 121.4% under soaked condition, with polypropylene fibres of 0.5%, 1%, 1.5% and 2% (by weight) respectively. Similarly, for coir fibre reinforced sand the improvement in CBR values were found to be 62.3%, 92.8%, and 114.3% under soaked condition, and 62.5%, 93.4% and 100% under unsoaked condition, with fibre content of 0.5%, 1% and 1.5% (by weight) respectively. Statistical analysis of triaxial tests results is carried out to estimate the Influence of fibre properties (i.e. weight fraction, aspect ratio, skin friction and modulus of elasticity), soil parameters (i.e. particle size, shape and gradation and, degree of compaction) and confining stress on the strength of fibre reinforced soils. A new mathematical model is developed to predict the shear strength of randomly distributed fibre reinforced soils. The model predicts well the strength of both poorly graded and well graded cohesionless soils for given soil-fibre parameters and stress conditions. The model can also be used for c - $ soils. The validity of the proposed model has been verified by comparing experimental test data and theoretically predicted strength envelopes and also comparing the published results. The findings of the present study are useful in understanding the behaviour of randomly distributed discrete fibre reinforced soils. They have practical significance as a ground improvement technique, with respect to embankment, subgrade/sub-base and other such problems. The proposed model is useful in estimating the strength of RDF