Measured and Predicted Durability and Mechanical Properties of Frozen-Thawed Fine Soils

An efficient experiment has been aimed in engineering fields for many years, which is provided through minimum number of tests and maximum amount of applicable results. Frozen soil is conceived to be an attractive subject for optimization processes. To reach an optimized design of experiment, a statistical method known as response surface method was used to largely reduce the number of tests and formulate a reasonable model for predicting the triaxial compressive strength and electrical resistivity. Considering the potential benefits of response surface method, a variety of effective factors including minimum freezing temperature during a cycle, number of freeze and thaw cycles and duration of freezing and/or thawing were investigated in order to have an overall outlook on the behavior of sandy lean clay subjected to different thermal regimes resembling natural conditions. In this study, triaxial compressive tests under confining pressures of 30 kPa and 90 kPa together with electrical resistivity tests were performed. In total, 38 treatments were performed under closed system to develop and verify models depending on mentioned factors, by which the mechanical behavior and durability of frozen soil could be reasonably estimated. The repeatability and relative errors of the derived models were also evaluated. It was shown that the derived statistical models were able to quantify the level of significance of each single input factor and their interactions with modeled properties, which can simplify the test protocol required to find the most destructive parameter and thermal regime. Minimum temperature and number of cycles were found to have the most significant effects on mechanical properties and durability, respectively.