Multi-Dimensional Frost Heave Modeling With Sp Porosity Growth Function

Abstract

This dissertation presents a multi-dimensional frost-heave modeling with coupled heat transfer, moisture transfer, and mechanical analysis. A series of laboratory frost-heave tests was conducted to determine segregation potential (SP) values using the effect of cooling rate and overburden pressure in two different freezing modes. Regardless of the freezing mode, consistent SP values were obtained at the formation of the final ice lens. Continuous heave and water-intake measurements made it possible to determine the time at the formation of the final ice lens. The SP porosity growth function was developed using simulations of the growing ice lens and frozen fringe. The developed frost-heave model was verified by laboratory frost-heave tests in one dimension. The simulated temperature distribution and amount of heave were in good agreement with experimental values. The SP porosity growth function was then expanded to two dimensions to simulate the soil-pipeline interaction of an experimental buried chilled pipeline constructed in Fairbanks, Alaska in the early 2000s. A two-dimensional frost-heave simulation was conducted at the free-field area, where the influence of pipeline resistance in frozen ground was negligible. This model, which considers the effect of frozen soil creep on stress distribution due to temperature variation, analyzed the influence of stress fields on soil frost-heave susceptibility and deformation. Simulations of pipe displacement were conducted for two cases, with and without the use of the long-term creep characteristics of frozen soils. Using the long-term creep characteristics, the simulated result agreed well with the observed value, differing by only a few percentage points. However, without using long-term creep characteristics, the simulated pipe heave was approximately 75% of the observed heave because of an unrealistic stress buildup. Finally, the SP porosity growth function was expanded to predict soil-pipeline interaction around a frozen-unfrozen boundary. Temperature distribution was successfully predicted in both the pre-frozen soil and the unfrozen zones, as well as at the time when differential pipeline movement started. The developed three-dimensional frost-heave model could predict pipe movement and induced bending due to differential frost heave for a 20-year period.