Resumen de Study of ice content and hydro-mechanical behaviour of frozen soils
The hydro-mechanical behaviour of soils during freezing and thawing is a coupled multi-physics process that has important practical applications. To cite but a few examples, on artificial ground freezing –to provide structural support and exclude the groundwater from underground constructions–, on seasonal variations of permafrost soils and their consequences on infrastructures, on geocomposite capillary barriers used to reduce frost heave in soils, and on engineered barriers subjected to freezing and thawing processes. Despite this importance, ice content and its migration process, as well as the coupled hydro-mechanical behaviour have not been extensively studied in partially saturated frozen soils. In these soils, ice formation increases the number of phases and the complexity of the experimental studies.
In this thesis, the unfrozen liquid saturation and the hydro-mechanical behaviour of two frozen soils (Barcelona clayey silt and fine sand) have been investigated at different initial degrees of saturation. A series of experimental setups have been designed and built to perform the tests (oedometer cell installed in a freezing/thawing chamber to control the temperature up to -15°C; setup for freezing cylindrical samples in an electrical resistivity tomography ERT cell with 16 lateral electrodes).
Different methods have been considered and several models proposed in the thesis to indirectly determine the unfrozen liquid saturation of compacted soils at different temperatures, porosities and initial degrees of saturation. The methods rely on measurements during transient freezing of the bulk electrical conductivity EC, relative dielectric permittivity, and thermal conductivity. In the case of EC tests, the unfrozen liquid saturation results have been interpreted with a modified Archie’s law and compared to the electrical conductivity of the pore liquid. The soil freezing retention curve (unfrozen liquid saturation vs temperature) has been also estimated by combining the Clausius-Clapeyron equation with water retention data on drying, as well as with a capillary bundle model with mercury intrusion porosimetry MIP results. A very good agreement has been found between the proposed approach using bulk EC measurements and the water retention and porosimetry results, which validated the proposed models.
The EC model has been used to interpret the transient freezing of cylindrical homogeneous samples that have been exposed to very low temperatures at its central axis (-15°C) and with adiabatic conditions imposed at the external boundaries. A 2D ERT setup is used to monitor the phase change and the migration of pore liquid. Reconstructed maps of EC have been translated into images of temperature and unfrozen liquid saturation at different elapsed times. The reconstructed temperature field showed good agreement with direct temperature measurements using inserted thermocouples. The ERT tests have been also performed on homogeneous samples with inclusions (low electrical conductive or high electrical conductive zones), which indicated that the ERT system could perfectly detect the effects of these anomalies on a transient freezing process.
The volume change behaviour on freezing/thawing has been investigated on saturated samples using temperature controlled oedometer cells under different constant vertical stresses. A small irreversible compression has been systematically detected after a freezing/thawing cycle on different soils, which depended on the current stress state. The microstructural changes of the compacted samples after freezing/thawing paths have been also explored using MIP and field emission scanning electron microscopy FESEM observations, which have been analysed using PCAS image processing software. MIP results have indicated that the freezing/thawing process might have decreased the macropore volume (between aggregates) and enlarge some micropore volume.
