Abstract
An experimental study on the behaviour of unreinforced, fibre-reinforced, and lime-stabilised residually derived soil for the subgrade layer in the laboratory to simulate the field condition is presented in this thesis. The soil composite was compacted in a model test box (representing a model road layer) using a steel-braced wooden box with the length, width and height of 460 mm, 410 mm and 1.0 m, respectively. A subgrade layer was prepared to a thickness of 650 mm and subbase of 200 mm. 150 mm from the uppermost part of the subgrade layer was reinforced with varying fibre contents, fibre type and hydrated lime; the remaining 500 mm thickness was unchanged, and then, the granular subbase was compacted on top of the subgrade layer.
A series of model plate load tests were conducted to investigate an unpaved road layer's load–settlement behaviour on fibre-reinforced and lime-stabilised soil with varying fibre contents and fibre type. Due to the plate size (150 mm) and to prevent the boundary effects from the model test box, a dynamic load test was conducted at 1 KN incremental loading. The diameter of the model test box was considered large enough to prevent or reduce possible boundary effects. LVDT were connected to a data acquisition system, and the results were recorded throughout the loading to check the buckling of the test box. At 75 mm settlement, where the fibres (1.8 % and 3.0 % strand and discrete fibre, respectively) have dramatically improved static loading, the resulting reinforced and unreinforced test sections were stabilised with 8 % lime demand for dynamic loading test. The results show that 1.8 % of discrete fibre-reinforced model test sections performed better under dynamic loading than other reinforced and unreinforced model sections. Also, the lime-stabilised model section performs better than the fibre-reinforced and unreinforced sections except at 2.5 mm settlement. Furthermore, a settlement resistance benefit for the long-term performance of settlement behaviour of fibre-reinforced, lime-stabilised, and unreinforced pavement models at a high settlement was established. The reinforced and lime-stabilised model section exhibits a better settlement resistance than the unreinforced and fibre-lime-stabilised model section, indicating that fibre is as effective as a subgrade reinforcement under a long cycle loading.
The California bearing ratio (CBR) is considered the most important parameter in evaluating the subgrade strength in pavement layers. As CBR does not adequately represent the in-situ condition, a new DCP was developed, and CBR values were derived using a correlation equation
iv | P a g e
per ASTM D6951. In this current research, the modified DCP cone tip with the geometry of an expanded cone of an angle of 60 degrees, cone base diameter of 35 mm and driving rod of 25 mm was changed into a flat surface (bar) with the geometry of 25 mm diameter and 20 mm thickness. At the same time, the conventional DCP cone tip with a geometry of an angle of 60 degrees, cone base diameter of 20 mm and driving rod of 16 mm was changed into a flat surface (bar) with a geometry of 16 mm diameter and 20 mm thickness to mimic the conventional CBR test for a fibre-reinforced subgrade layer in a model road layer and modified AASHTO mould. A test protocol consisting of a Dynamic Cone Penetrometer test (DCPT) and California Bearing Ratio (CBR) tests were implemented on a model road layer reinforced with polypropylene fibres (strand and discrete fibre), CBR mould (150 mm x 150 mm) and modified AASHTO mould (150 mm x 300 mm). The effects of DCP on the modelled road layers with strand and discrete fibre, hydrated lime, different hammer weights (2.5 kg, 4.6 kg and 8.0 kg), and different DCP types (modified and conventional) with a flat and cone surface were investigated. The compaction energy of the compaction done in the standard mould with a fibre-reinforced model is 3426.38 and 1440.47 kJ for CBR and modified AASHTO moulds, respectively. At the same time, the compaction energy of the compaction done for the unreinforced model was 2936.90 and 1234.68 kJ for CBR and modified AASHTO mould, respectively. The specific work done per blow for DCP with cone tip geometry and hammer weight of 2.5 kg, 4.6 kg, and 8 kg were 564.08, 1037.90, and 1505.04 kJ for conventional DCP, respectively and 1.83, 3.37, and 5.86 kJ for modified DCP respectively. Also, the specific work done per blow for DCP with flat tip geometry and hammer weight of 2.5 kg, 4.6 kg, and 8 kg were 7.46, 13.73, and 23.88 kJ for conventional DCP, respectively and 3.46, 6.38, and 11.09 kJ for modified DCP respectively. The test results of DCP were expressed in terms of Dynamic Cone Penetration Index (mm/blow), defined as the penetration depth of the cone per hammer blow and recorded along with the depth profile. The results obtained with the two DCPs with bar tips were compared with that of the cone tip configuration. A decrease in DCPI value was observed for the reinforced test sections compared to the unreinforced test section and for fibre lime-reinforced soil at the fibre content of 2.4 and 3.0 % strand and discrete fibre, respectively. With optimum lime demand of 8 %, the DCPI increased for lime-stabilised subgrade soil but decreased with fibre lime-reinforced soil. DCP results detected a transition zone and significant change in the strength of the unpaved test section along with penetration depth with the hammer weight of 2.5
v | P a g e
kg and modified DCP. Higher penetration resistance offered by the fibre contributes more to the performance of the unpaved road model section.
A set of tests consisting of a matric suction test (using a filter paper approach) and the indirect tensile strength test (ITS) on a fibre-reinforced soil compacted in a circular metal ring (120 mm x 60 mm) resulted in an aspect ratio of 2 were investigated to determine the suction effects and the tensile properties of fibre reinforced soil. Filter-paper method for matric suction measurement was employed per ASTM D5298. While for the indirect tensile strength test, the test was conducted per ASTM D 6931-17, where three trials were performed for each test, and the average was taken for accuracy. A calibrated wetting curve was used to evaluate matric suctions. Whatman 42 100 mm diameter and 120 mm diameter filter papers were used for the matric suction test. Matric suction recorded high suction values due to the presence of fibre, which increased with fibre content for different fibre types. It was shown that matric suction increased with fibre content and initial moisture content due to the film of water on its surface, which improved its ability to mix well with the soil and thereby reduced the water content compared to the unreinforced soil. The peak tensile strength of fibre-reinforced soil increases with an increase in fibre content up to an optimum value. The optimum fibre content was 1.8 % and 2.4 % for strand and discrete fibre, respectively. Finally, residually derived soil formation, which is very common in semi-arid environments, needs alternative improvement methods which are sustainably inclined. Polypropylene fibre (strand and discrete fibre) was investigated in large quantities to enhance the strength of the subgrade soil. The results from the ITS test, CBR, matric suction, DCP, and plate loading test show that fibre can improve the subgrade soil when used in large quantities, but if it is used between 1.8 % and 2.4 % fibre contents and that the results from the model test box are a good representation of the field condition. However, strand fibre performs better than discrete fibre between 1.2 and 2.4 % fibre content. It drastically reduced at 3.0 % fibre content due to the bearing pressure sustained and the equivalent settlement at that bearing pressure. Fibre has no benefits at lower loading cycles when used in large quantities. This phenomenon might occur due to the relative slippage between fibres and soil particles being very limited due to the soil particles' low compaction.