THE INFLUENCE OF SOIL MASS FRACTAL DIMENSION ON THE PROPERTIES OF CEMENT STABILIZED SOIL

In roadway construction, the cost of soil-cement stabilization per unit area is significantly influenced by the binder content, hence the need to optimise cement usage. This research work characterises the influence of soil fractal geometry on properties of cement-stabilized soil and strives to determine a correlation between mechanical proprieties of cement-stabilized soil and the mass fractal dimension Dm indicated by particle size distribution (PSD) of aggregate mixtures. Since strength development in cemented soil relies not only on cement content but also on soil PSD, this research investigates the possibility of reducing cement content by changing the mass fractal distribution (MFD) of soil without compromising on strength, reduced permeability and compressibility. A quantitative inductive research approach will be adopted to achieve the aim of this research. This approach seeks to create implications from the data set collected through a series of laboratory investigations and data analysis, to identify patterns and connections to build a theory. In the MPhil phase of this research, an extensive background study was conducted, and several pieces of literature were reviewed. A methodology for formulating the artificial soil samples was adopted from a research by Giménez, et al. (1997). Trial soil mixtures were formed from a combination of fractions of the same soil. A series of trial experiments were conducted to test the power-law mathematical model for determining soil mass fractal dimension proposed by Tyler and Wheatcraft (1992), to verify the correlation between soil mass fractal distribution (MFD) and soil particle size distribution (PSD). Dry sieving tests were performed on formulated soil mixtures containing different particle size distributions. During the PhD phase of this research, laboratory tests will be performed on formulated soil mixtures, with various particle size distributions. The mass fractal dimension Dm of each mix will be determined from sieve analysis data using the power-law model, (M(r