dc.description.abstract | Soil has very important functions such as to support buildings, railway lines, pipelines,
dams and roads. However, good site characterisation or accurate geotechnical
investigations will enhance a good design of the foundation to support the structure.
Sound geotechnical investigation requires reliable soil data in the area of interest so as
to achieve accurate information about the site. A combination of in-situ testing, site
description, use of ancillary data (such as aerial photographs, remote sensing images,
geologic maps, etc) , soil and rock sampling for further laboratory tests is used during
investigations.
In order to understand and interpret the soil properties of the site before any design is
made, a 3D view is essential. This is because it enhances visualization of geotechnical
properties in three dimensions and facilitates interpretation of the underground
information. The major problem in geotechnical presentation is uncertainties due to
prediction of soil properties in un-sampled areas. Field observations are usually spaced
to minimize cost of drilling boreholes at each and every meter, hence; interpolation
between these widely-spaced data points is done. In this research, geostatistics with R
computing software was used in spatial interpolation of soil properties in un-sampled
areas.
A three-dimensional model for geotechnical investigation was developed using
observations from seven boreholes drilled to a depth of 14m in the study area. The soil
properties were obtained from both in-situ and laboratory tests. R code was developed
during the study where regression and Kriging methods (hybrid interpolation technique)
were used to determine spatial variation of geotechnical soil properties such as
plasticity index, SPT blow counts etc. The nugget, sill and range of the experimental
varioqrarn model were determined at 0.0037, 0.0218 and 2000m respectively for
plasticity index.
The spatial interpolation model for discrete geotechnical investigations with the
computer R code were used to produce 3D visualization of soil properties in the
boreholes. The SPT blow counts predicted in the entire study area varied from 0-50 N
values with high resistance as from 7m deep. Soil types encountered were mainly
marls, clayish sandy deposits.
The developed 3D model can be used to improve geotechnical investigations where the
distribution of the measured soil properties for example SPT (N) counts within the area
of interest is visualized. This will greatly intensify fairly accurate and economical
geotechnical design since the cost of data collection, analysis and mapping will be
reduced by increasing the spacing of the boreholes. | en |