The design of pile foundations in conventional geotechnical engineering practice is based on the soil mechanics principles for saturated soils. These approaches are also extended to pile foundations that are placed totally or partially above the ground water table (i.e., vadose zone), where the soil is typically in a state of unsaturated condition. Such approaches lead to unrealistic estimations of the load carrying capacity and the settlement behavior of pile foundations. Some studies were undertaken in recent years to understand the influence of the matric suction towards the bearing capacity of model pile foundations placed in unsaturated fine-grained and coarse-grained soils. The conventional   and methods were modified to interpret the contribution of shaft carrying capacity of single piles in fine-grained soils (e.g., Vanapalli and Taylan 2011, Vanapalli and Taylan 2012). Also, the conventional method has been used to understand the contribution of matric suction towards the shaft resistance in unsaturated sands (Vanapalli et al. 2010). One of the key objectives of the present research study is directed to determine the contribution of matric suction towards the bearing capacity and settlement behavior of model single pile foundations in unsaturated sands. A series of single model pile load tests were performed in a laboratory environment to study the contribution of the matric suction towards the total, shaft, and base bearing capacity of the model piles with three different diameters (i.e., 38.30, 31.75, and 19.25 mm) in two unsaturated sands (i.e., a clean commercial sand and a super fine sand). Hanging column method (i.e., plexi glass water container) was used to control the matric suction values in the compacted sands in the test tank by varying the water table. The results of the testing programs indicate the significant contribution of the matric suction towards the bearing capacity of single model piles (i.e., 2 to 2.5 times of base bearing capacity and 5 times of shaft bearing capacity under unsaturated conditions in comparison with saturated condition). The test results were interpreted successfully by modifying the conventional methods for estimating the pile shaft bearing capacity (i.e., β method) and base bearing capacity (i.e., Terzaghi 1943, Hansen 1970 and Janbu 1976). In addition, semi-empirical methods were proposed for predicting the bearing capacity of single model piles using the effective shear strength parameters (i.e., c' and ϕ') and the soil-water characteristic curve (SWCC). There is a good agreement between the measured and the predicted bearing capacity of single model piles using the semi-empirical models proposed in this study. In addition, numerical investigations were undertaken using the commercial finite element analysis program SIGMA/W (Geostudio 2007) to simulate the load-displacement (i.e., p-δ) behavior of the single model piles for the two sands (i.e., clean commercial sand and super fine sand) under saturated and unsaturated conditions. An elastic-perfectly plastic Mohr-Coulomb model that takes into account the influence of the matric suction was used to simulate the load-displacement (i.e., p-δ) behavior. The numerical approach proposed in this thesis is simple and only requires the information of the effective shear strength parameters (i.e., c' and ϕ'), the elastic modulus (i.e., Esat) under saturated conditions, the soil-water characteristic curve (SWCC), and the distribution of the matric suction with respect to depth. The approaches proposed in this thesis can be extended to determine the in-situ load carrying capacity of single piles and also simulate the load-displacement (i.e., p-δ) behavior. The studies presented in this thesis are promising and encouraging to study their validity in-situ conditions. Such studies will be valuable to implement the mechanics of unsaturated soils into geotechnical engineering practice.

This study explored the available adfreeze data published in literature and the techniques used to obtain it. Two methods were selected and modified to complete series of adfreeze bond test. A model pile pull-out method consisting of pulling a pile out a large specimen of soil was the first method used. The second method was modified from an interface shearing apparatus developed by Dr. Fakharian and Dr. Evgin at the University of Ottawa in 1996 and allowed preparing, freezing and testing the specimen in place. The material and soil tested for this study were provided by EXP Services Inc. The model pile, a galvanized HSS 114.3 x 8.6 section, is commonly used to install solar panels. Soil was taken from a future solar farm site in proximity to Cornwall, Ontario. The study had for objective to develop a low cost adfreeze laboratory testing method. Limitations of the technics and apparatus used were observed. While the results of a model pile pull-out test compared to previous data publish by Parameswaran (1978), the interface shear series of test presented more limitations. The interface shearing method has been previously study by Ladanyi and Thériault (1990). Issues with the interface shear method due to the water content of the soil as well as the range of normal stress applied to the specimen both during testing and freezing. The data obtained was inconclusive and the method will be studied in future research program. This studied approach the adfreeze testing with new improvement. The main contribution of this study is the data obtained by measuring and observing adfreeze of ice poor sand with varying water content. The measurements allowed to study the effect that increasing water content has on the interface bond strength. The modifications made to interface shear apparatus are also major new contribution provided by this research. The apparatus was converted in a small freezer chamber using insulation panel and vortex tubes. Which was used to freeze the specimen in the testing chamber and testing adfreeze in place without handling the shear box arrangement.

The design of pile foundations in conventional geotechnical engineering practice is based on the soil mechanics principles for saturated soils. These approaches are also extended to pile foundations that are placed totally or partially above the ground water table (i.e., vadose zone), where the soil is typically in a state of unsaturated condition. Such approaches lead to unrealistic estimations of the load carrying capacity and the settlement behavior of pile foundations. Some studies were undertaken in recent years to understand the influence of the matric suction towards the bearing capacity of model pile foundations placed in unsaturated fine-grained and coarse-grained soils. The conventional   and methods were modified to interpret the contribution of shaft carrying capacity of single piles in fine-grained soils (e.g., Vanapalli and Taylan 2011, Vanapalli and Taylan 2012). Also, the conventional method has been used to understand the contribution of matric suction towards the shaft resistance in unsaturated sands (Vanapalli et al. 2010). One of the key objectives of the present research study is directed to determine the contribution of matric suction towards the bearing capacity and settlement behavior of model single pile foundations in unsaturated sands. A series of single model pile load tests were performed in a laboratory environment to study the contribution of the matric suction towards the total, shaft, and base bearing capacity of the model piles with three different diameters (i.e., 38.30, 31.75, and 19.25 mm) in two unsaturated sands (i.e., a clean commercial sand and a super fine sand). Hanging column method (i.e., plexi glass water container) was used to control the matric suction values in the compacted sands in the test tank by varying the water table. The results of the testing programs indicate the significant contribution of the matric suction towards the bearing capacity of single model piles (i.e., 2 to 2.5 times of base bearing capacity and 5 times of shaft bearing capacity under unsaturated conditions in comparison with saturated condition). The test results were interpreted successfully by modifying the conventional methods for estimating the pile shaft bearing capacity (i.e., β method) and base bearing capacity (i.e., Terzaghi 1943, Hansen 1970 and Janbu 1976). In addition, semi-empirical methods were proposed for predicting the bearing capacity of single model piles using the effective shear strength parameters (i.e., c' and ϕ') and the soil-water characteristic curve (SWCC). There is a good agreement between the measured and the predicted bearing capacity of single model piles using the semi-empirical models proposed in this study. In addition, numerical investigations were undertaken using the commercial finite element analysis program SIGMA/W (Geostudio 2007) to simulate the load-displacement (i.e., p-δ) behavior of the single model piles for the two sands (i.e., clean commercial sand and super fine sand) under saturated and unsaturated conditions. An elastic-perfectly plastic Mohr-Coulomb model that takes into account the influence of the matric suction was used to simulate the load-displacement (i.e., p-δ) behavior. The numerical approach proposed in this thesis is simple and only requires the information of the effective shear strength parameters (i.e., c' and ϕ'), the elastic modulus (i.e., Esat) under saturated conditions, the soil-water characteristic curve (SWCC), and the distribution of the matric suction with respect to depth. The approaches proposed in this thesis can be extended to determine the in-situ load carrying capacity of single piles and also simulate the load-displacement (i.e., p-δ) behavior. The studies presented in this thesis are promising and encouraging to study their validity in-situ conditions. Such studies will be valuable to implement the mechanics of unsaturated soils into geotechnical engineering practice.