Étude expérimentale du dégel d'un limon argileux

Curiously, the thawing of frozen soils containing ice lenses has only been the subject of rare fundamental studies even though it is the cause of the greatest damage to constructions. The reduction of bearing capacity caused by the thawing of the ice lenses is a destructive phenomenon, and thus costly, which concerns roadway and railway infrastructures, as well as the melting of mountain permafrost by global warming. The present research, especially experimental, has tried to explain how ice lenses melt and how the water produced by the melting acts on soil properties, in particular on their deformability. Numerous freezing and thawing tests have been carried out in a testing apparatus including : A mould containing a specimen 150 mm in diameter and 300 mm in height. It is slightly conical in order to reduce the friction against its walls due to swelling. Numerous gauges placed on the sides and in the specimen in order to measure temperature, unfrozen water content and suction. Three cryostats, which control temperature at the head of the specimen (positive and negative), at its base (always positive) and outside of the thermal insulation placed against the mould. A micro-camera (endoscope), which moves in a translucent tube placed along the axis of the specimen, which enables animations of the growth and melting of ice lenses to be made. A press to enable loading and unloading cycles to be applied to the specimen. An X ray device using lead shot placed in the specimen, which enables the measurement of deformation in the entire specimen during the freezing and thawing cycles. All of the tests were carried out on one very frost–susceptible silt. Their duration of approximately two months excluded the possibility of carrying out tests on several types of frost–susceptible soils. Some numerical simulations permitted the verification of the thermal behaviour of the test apparatus. Then, previous freezing and thawing tests carried out at full scale on road pavements, carried out in a large test pit, were reinterpreted in order to obtain improved information on the deformability (resilient moduli) of the frost–susceptible infrastructure which was made up of a silt similar to that tested in the laboratory tests described above. The results of the laboratory tests and the reinterpretation of the full–scale measurements on road pavements were used for two very different practical applications : the design of roadway and railway pavements by quantitative methods using, in particular, resilient moduli and the formation of mountain permafrost and its thawing due to global warming. In the general field of the physical phenomenon of freezing and thawing of frost–susceptible soils, the very elaborate experiments permitted the measurement, with precision and reliability, of numerous parameters which are involved in the phenomena of freezing and thawing of fine-grained soils. However, most of these phenomena were already known and their parameters have been determined rather well by numerous experiments. This research has, nevertheless, permitted the quantification of certain parameters a bit better. The greatest contribution to the understanding of these phenomena has been the visualisation of the growth and the thawing of ice lenses inside the specimen using an endoscope, which has never been carried out previously. The resulting animations show extremely well how the ice lenses form and melt in a virtually natural soil. In the field of freezing and thawing design for roads and railways, the research has shown that the use of a thawing resilient modulus combined with modern numerical design methods is completely possible. This research has provided several values for this modulus : for the silt used in the full–scale test pit, for various frost–susceptible soils, and, by a more elaborate method, for the silt used in this research. Also concerning this field, it has shown how the resilient modulus may be used in modern numerical methods and how to determine it in the laboratory. The results of this research will thus be very useful for the work of the Swiss committee which will very soon address the revision of Swiss standards for the design of pavements for roads and railways. In the field of the construction of roads and railways, this research has confirmed that the water, which comes from the melting of ice lenses, flows toward the soil zone from where it has been extracted then transported toward the freezing front. Any drainage of the formation level is thus of no practical use. Finally, in the field of melting of permafrost due to global warming and the initiation of debris flows in mountain permafrost, this research has shown, first of all, the importance of the capillary suction regime in the formation of mountain permafrost, which has never been well explained until now. It has also shown how the use of sophisticated numerical models in the prediction of the melting of mountain permafrost and of the initiation of debris flows is possible. Such use, however, is difficult and should not be considered as a practical tool for common use in the prevention of these dangerous phenomena. This research has also provided a practical method for evaluating the degree of supersaturation of permafrost.