Predicting Moisture - Density Relationship from the Index Properties of Lateritic Soils using Regression Analysis: A Case study in Durame Town

Abstract

Modification of soils at a construction site to improve engineering properties is essential in many applications (e.g. highways, subgrades, earth dams, building foundations and earth-retaining structures) to improve their engineering properties. Mechanical compaction is used as a practical means of achieving the desired strength, compressibility and hydraulic conductivity characteristics of soils. The compaction curve and associated characteristics namely maximum dry unit weight and optimum moisture content are determined in the laboratory using compaction test procedures. The test not only requires significant time and effort to carry out, but also require relatively large quantity of soil. Hence, there is a necessity to predict compaction characteristics of soils with the help of correlation analysis with the index properties of soils, which can be determined easily. The study focused on the development of the statistical model between moisture density relation and index properties of lateritic soil specifically located in Durame town. For this purpose, 30 samples collected, and a series of laboratory tests conducted. The parameters considered for this study are Atterberg's limits, Grain size, Specific gravity, and Compaction tests. The test procedures based on American Society of Testing and Materials laboratory test standards. Results indicated, the soil type found in Durame town composed of inorganic highly plastic latritic red clay soil, and inorganic silts of high plasticity. In the analysis part, both the MS excel spreadsheet and the SPSS software had used for the scatter plot, correlation, and regression analysis. Based on both single and multiple linear regression analysis, relatively good correlation was obtained by combining plastic limit and liquid limit with maximum dry density and optimum moisture content of lateritic soil. From the correlation analysis, the equations developed are MDD = 1.765 - 0.004LL with coefficient of determinant of 0.831, and OMC = 13.798 + 0.278LL with coefficient of determinant 0.884 for single linear regression, and MDD = 1.758 - 0.002LL - 0.005PL with coefficient of determinant 0.887, OMC = 13.924 + 0.225LL + 0.1PL with coefficient of determinant of 0.908 for multiple linear regression, respectively. It has been shown that these models will be useful for a preliminary design of earthwork projects which involve lateritic soils in Durame town.