Frost Heave: A Semi-Empirical Model Based on Field Data

摘要

In cold regions, as the soil temperature drops below the freezing point, available water starts to freeze and form ice lenses. The growth of ice lenses depends on many variables, including temperature, soil type, capillarity, availability of water, proximity of the ground water table, and applied load or overburden pressure. Ice growth causes heave in the soil, which may precipitate extensive damage to structures such as pavements, shoulders, utility lines, utility poles, and even unprotected foundations. Most existing mechanistic-empirical frost heave prediction models are based on two theories, capillary (primary frost heave) and frozen fringe (secondary frost heave). Unfortunately, none of the models have received universal acceptance. The capillary theory has some limitations, which result in inaccurate prediction of frost heave. The frozen fringe theory involves complex models that require numerous input variables, which are often unavailable or expensive to collect. This paper addresses the modifications that are made to one of the secondary frost heave models. The modified model uses simplified input data and a statistical frost depth model that is developed during the study using measured frost depth data from the state of Michigan. The use of the statistical frost depth model simplifies the solution and improves the accuracy of the results. Finally, the modified model is evaluated using the pavement and shoulder frost heave data collected in the state of Michigan.