UTHM Institutional Repository

Effects of differently stabilised lateritic soils on its geotechnical properties and dust generations

Lim, Sin Mei (2015) Effects of differently stabilised lateritic soils on its geotechnical properties and dust generations. PhD thesis, Universiti Tun Hussein Onn Malaysia.

[img]
Preview
PDF
lim_sin_mei.pdf

Download (5MB)

Abstract

Well designed roads are the vital socio-economic pathways of a nation. However, construction of unpaved roads in rural areas are often hindered by geographical limitation and are deemed costly and energy inefficient. This research assesses the soil stabiliser performance through the evaluation of the geotechnical properties of the unstructured and structured soil using frontline laboratory testing on soils collected from 2 case study sites in Johor, Malaysia. There were treated with 2, 4, and 6% of a powder stabiliser (PPS) and 0.096, 0.1 19, and 0.144% of a liquid stabiliser (PLS) and individually cured under controlled conditions to 7, 14 and 28 days. The outcomes confirmed that that with the use of reliable stabilisers and ageing, the strength and stiffness behaviour of the structured soils improved significantly as desired. The 44 in-house research datasets collected were complemented with nearly 300 sets of published data to establish correlations through simple and multiple regression analysis using the tools made available in MINITAB 17. Various multiple correlation equations presented in this thesis bore high coefficient of significance (pvalue < 0.05); conclusively demonstrating good prediction. A new soil stabilisation model: [S] [MI = [L] where [ ] represents the appropriate matrix formulation for [S], properties of the unstructured soil; [MI, stabilised soil properties and [L], desired designed properties for the structured (treated) soils through the conceptual use of Artificial Neural Network (ANN) is proposed and demonstrated in this research. "NeuroShell Predictor" was the engine used in the ANN training and testing. The reliable correlations developed concurrently in this research formed the hidden neuron layer of the ANN. Performance and acceptance of the ANN is evaluated using R-squared ( R >~ 0.9) and the accuracy of the ANN models were measured using the Root Mean Square Error (RMSE). This thesis also adopted the holistic approach to emphasise that soil stabilisation must globally and concurrently encompass satisfactorily the two extreme scenarios of soil softening under wet conditions and hazardous dust formation during extreme dry conditions. Hence, an innovative dust assessment technique was developed to assess the improvement in the "soil erosion index" with stabilisation.

Item Type: Thesis (PhD)
Uncontrolled Keywords: artificial neural networks; California bearing ratio; geotechnical properties; soil erosion index; soil stabilisation
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA703-711 Engineering geology. Rock mechanics. Soil mechanics.
Depositing User: Normajihan Abd. Rahman
Date Deposited: 20 Jan 2016 08:34
Last Modified: 20 Jan 2016 08:34
URI: http://eprints.uthm.edu.my/id/eprint/7486
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year