The coefficient of earth pressure at rest (K
0) is a key parameter pertaining to many problems in geotechnical engineering in which the condition of zero lateral movement applies. A range of applications requires the knowledge of K
0 during virgin loading and/or unloading at varying overconsolidation ratios (OCRs). The determination of K
0 is mainly based on empirical models incorporating soil’s friction angle (∅’) and OCR or through direct laboratory testing. Measurement of K
0, however, is challenging as it requires especially designed devices, for which tactile pressure sensors (TPS) may provide an easy-to-fabricate, low-cost measuring system. Although it is well known that particle morphology could remarkably affect the micro and macro parameters including K
0, the role of morphology has not been incorporated into existing models to develop more robust models for granular soils. Furthermore, the selection of the representative sample size (RSS) for morphology characterization mostly relies on an intuitive selection or based on methods that are mainly subjective which may or may not result in capturing the representative morphology. Two systematic algorithms for objectively determining the RSS required for soil morphology characterization that can ensure either the similarity of mean values or probability density functions (PDFs) between an RSS and its population were developed. Sensitivity analysis of both algorithms showed that the developed algorithms are capable of providing reliable values of RSS. Applying the developed algorithms to two-and three-dimensional (2D and 3D) imaging data of standard sands with a wide range of angularity indicated that the RSS required to obtain mean particle morphology ranges from 10 to 160 while capturing the PDF of morphological descriptors requires larger RSS in the range of 350 to 600. These ranges are 2 to 10 times smaller than the RSS values obtained from the currently available methods, thereby resulting in significant saving of time and effort while obtaining similar morphologies. The concept of RSS was used to fairly and systematically evaluate the efficacy of morphology characterization using digital imaging devices as well as a light microscope (LM) in comparison to µCT results. The results indicated that since using RSS concept for selecting the number of imaged particles ensures characterizing the variability in shape of particles, the PDFs and mean values obtained from the LM have acceptable similarity to µCT results. Moreover, the of use digital imaging devices for determining very fine to medium fine morphological descriptors is not suggested until significant improvement in their camera resolution and especially computational algorithms in the built-in software has been attained. The LM is readily available in most cases and RSS ranges obtained are feasible to be imaged using an LM, thereby morphology of soils can be economically and conveniently determined. An extensive calibration campaign was conducted to evaluate the performance of point-type TPS considering different factors such as hysteresis, long-term drift, and effects of soil particle size and density. It was found that the sensor is accurate at ratios of the diameter of the sensing area to mean soil particle size D
sens/D
50≥10.5 in poorly-graded soils. A reliable test apparatus for measuring K
0 during loading and unloading was developed by integrating two TPS into a rigid oedometer cell and it was used to systematically study the effect of particle morphology on the K
0 of five granular materials where the morphology of the test materials was precisely characterized based on the RSS concept. Robust empirical models incorporating soil OCR and particle morphology were developed for estimating K
0. Morphology descriptors measurable at mesoscale and macroscale levels, namely roundness and sphericity, were found to substantially affect the K
0 during both loading and unloading. Elongated particles with several protrusions on their surfaces (i.e., more angularity) provide more particle contacts and hence promote interlocking, leading to a stronger soil structure that requires less pushing sideways (fewer horizontal stresses) to prevent lateral movement and thus have smaller K
0. The current practice of determining K
0 using ∅’ and OCR cannot provide reliable estimates of K
0 in granular materials because of ignoring particle morphology effect.
| Date of Award | 2023 |
|---|
| Original language | English |
|---|
| Awarding Institution | - The Hong Kong University of Science and Technology
|
|---|
| Supervisor | Anthony LEUNG (Supervisor) |
|---|