AN ITERATIVE PROGRAM TO BACK ANALYZE GRAIN-SIZE DISTRIBUTION FROM A PREDETERMINED SOIL-WATER CHARACTERISTIC CURVE
MetadataShow full item record
Soil-water characteristic curve (SWCC) is the relationship between matric suction and water content. The SWCC is the most important property in unsaturated soil mechanics as it contains information on seepage, shear strength, volume change and heat flow. Numerous methods of predicting the unsaturated properties (SWCC) and unsaturated permeability based on basic soil parameters have been introduced to reduce the cost of and skill required for unsaturated soil testing. This research proposes another unsaturated soil estimation method that used a predetermined SWCC to estimate the grain-size distribution (GSD) using computer iteration. The study was limited to laboratory scale testing and coding work. The goal of this project is to construct a back analysis program to predict GSD from a predetermined SWCC. The specific objectives for this research were: (1) Determine correlation(s) to relate AEV to characteristic particle diameters for fine-grained soil, which were achieved by completing the following tasks: a. Conduct literature review to collect published GSD and SWCC data of fine-grained soils. b. Conduct linear regression analysis on the published GSD and SWCC data to determine correlation(s) between AEV and characteristic particle diameters of fine-grained soils. (2) Determine a suitable SWCC prediction method and construct the method as a Python program, which were achieved by completing the following tasks: a. Conducting literature review on the existing SWCC estimation methods. b. Evaluate the strengths and weaknesses to find the most suitable method for this research. c. Address the limitations presented in the chosen SWCC estimation methods that would affect the back analysis method. d. Conduct a sensitivity analysis on each of the input of the methods to understand the effect on the predicted result. (3) Construct a back analysis program and evaluate the capability of the program, which were achieved by completing the following tasks: a. Laboratory testing for necessary soil properties to use for the input of the model and to validate the prediction results: b. The back analyzed GSD was put through the SWCC estimation method. The predicted SWCC produced from the back analyzed GSD was compared with the predetermined SWCC to evaluate the capability of the back analysis program. Correlation(s) determined from Objective One were incorporated to the back analysis program to evaluate its capability on fine-grained soils. A new SWCC prediction method from GSD was created based on the theories of the Fredlund (2000) method. This method is termed the modified Fredlund method. The results showed that the modified Fredlund method can make reasonable SWCC prediction. A Monte-Carlo approach examining the variations to the GSD and the associated packing factor on the SWCC prediction are provided. Using this new prediction method as the basis for the back analysis program, it was possible to back-calculate a GSD from a given SWCC as well as a potential way to determine different combinations of GSD to produce the same SWCC.
DegreeMaster of Science (M.Sc.)
DepartmentCivil and Geological Engineering
CommitteeFleming, Ian; Smith, Laura; Fredlund, Murray; Elwood, David
SWCC, GSD, soil suction, back analysis