Testing of an artificially weakly bonded material has been carried out to help our understanding of the role of bonding in natural soils, such as tropical residual soils that usually bonded in nature and frequently exist in an unsaturated state. The use of artificial specimens allows reproducibility of bond strength, which is difficult to achieve with naturally bonded soils, such as residual soil. The artificial samples were prepared with void ratio of0.6 by mixing sand and kaolin before firing at 500°C for 5 hours. A new technique for producing looser samples was also developed using coriander seeds. These artificial bonded samples are reproducible and are not agedependent. The bond formed has many advantages over cement or other bonding agents that require a curing period and show a change in strength with time. In order to observe the interaction between sand and kaolin, the thin sections of the samples were prepared and investigated under a microscope. Microscopic studies showed that the fired kaolin forms bridges between the sand grains, as well as coating the grains. The loose sample showed a very small amount ofash residue left after firing at high temperature. The engineering behaviour of these artificially weakly bonded samples was investigated in triaxial tests. Destructured material (where the bonding was broken down) was also tested to provide a reference state for a comparison. Conventional drained and undrained triaxial compression tests were performed on artificially bonded and destructured samples in a saturated state. A series of constant water content tests were also carried out on artificially weakly bonded samples in an unsaturated state at various initial matric suctions. The saturated behaviour of destructured and bonded samples was studied in order to understand the influence of bond structure and provide a useful reference for the interpretation of the unsaturated tests. The destructured samples showed no significant peak in q/p' ratio in comparison with the bonded samples. All bonded and destructured samples sheared at lower stresses sustained higher values of stress ratio than those sheared at higher confining pressure. The effect of bonding was clearly seen when comparing the stress ratio values between bonded and destructured materials. The bonded samples achieved a higher stress ratio values if compared with . the destructured samples. However, with an increase in stress level, the bond strength decreases progressively at failure, therefore the peak in stress ratio dropped closer to that of the destructured material. A nonnalisation of the stress paths was perfonned in order to compare the results from different tests on bonded and destructured materials. The effect ofbonding was evident from the nonnalised stress paths. A series ofconstant water content tests were perfonned on unsaturated samples with suction ranging from zero up to 500kPa. The unsaturated samples were initially consolidated under mean net stresses of 50, 100 and 300kPa before shearing under constant water content conditions. The water retention curve (WRC) was detennined for a drying path. It showed a very steep change at suction between 2 - 4kPa before flattened when the degree of saturation, Sr dropped below 25%. During consolidation, samples experienced initial compression before levelled off with time. Samples with higher suction showed less compression in comparison to samples with lower suction. The matric suction for all samples, generally showed an initial decrease after applying a mean net stress, p-ua, but then levelled offwith time. The stress-strain curve for samples with zero and low suction sheared at lower stress showed a clear peak compared to samples sheared at higher stress which indicated more ductile behaviour. The volume change of the samples during shearing showed the influence ofmean net stress and suction. The peak strength surfaces plotted in q suction space were parallel and indicated some curvature. The curvature of the ultimate strength in shear strength - suction relationship showed that the contribution ofsuction to strength represented by friction angle, ~b was not constant. The initial and final bond yield surfaces were represented by a linear increase with mean net stress over the stress range investigated. As matric suctions increased higher, both bond yield surfaces dropped back close to yield surface for lower matric suctions. The critical state stress ratios Ma (mean net stress component) and Mb (matric suction component) were used to examine the critical state of the bonded material in unsaturated conditions. It was found that Ma is higher than the saturated critical state stress ratio, Ms (Le. cpa> cp), therefore, the assumption that cpa =<p' is not always valid. The changes in Ma and Mb can be related to three different phase of water retention behaviour. This pattern of behaviour is probably explained by the narrow range of pore size in the uniform bonded sand. A relationship between water content and matric suction was studied based on the WRC at each stage of suction equalisation, consolidation and shearing. The COLs or Continuously Disturbed Lines (representing the water content - suction relationship at ultimate state) were plotted. It was interesting to see that samples with initial low suction showed an increase in suction to reach the COL while samples with initial high suction showed a decrease to approach the COL. The decrease in suction during shearing for samples with high initial suction was accompanied by dilation behaviour in terms of volumetric strain. This behaviour is the opposite of what might be expected. In saturated soils, dilative behaviour would produce a drop in pore water pressure (increase in suction). This opposite behaviour expressed by unsaturated samples under shear emphasises the role of fabric in controlling the engineering behaviour of unsaturated soils.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:505091 |
| Date | January 2008 |
| Creators | Ali Rahman, Zulfahmi |
| Publisher | Durham University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://etheses.dur.ac.uk/1352/ |
Page generated in 0.0029 seconds