Final Report Summary - AREUS (Anisotropic Response of Unsaturated Soils: A Microstructural Approach)
A novel experimental technique based on image analysis has been developed to measure volume changes of soils subjected to drying-wetting cycles over a wide suction range. The technique uses saturated salt solutions to generate soil suctions corresponding to values of ambient relative humidity between 33 % and 97 %. Cubic soil specimens of 10 mm side are positioned inside a small cubic cell made of transparent acrylic and a peristaltic pump is used to force circulation of humid air (conditioned by the chosen salt solution) through the cell. Different soil specimens are simultaneously exposed to different values of relative humidity in separate cells positioned along a circumference and equidistant from a high-resolution macro-lens digital camera. The camera is fitted with a rotary device that allows taking constant-angle shots of the specimens. Dedicated image processing software is used in combination with the rosette gauge theory to define the planar strain field at different sample locations, including magnitude and orientation of principal strains.
The advantages of this experimental technique are summarised as follows:
a) the disturbance caused by specimen handling and the lengthy equalisation time following the manual measurement of the sample dimensions are avoided thanks to non-contact determination of strains via image analysis;
b) the testing time is highly reduced in comparison with conventional techniques because of forced air circulation and simultaneous testing of multiple (small) specimens at different values of relative humidity;
c) high-quality images with negligible geometrical distortion are recorded thanks to the use of a single lens reflex (SLR) photographic camera - equipped with an advanced photo System type-C (APS-C) sensor - in combination with a macro lens;
d) geometrical distortion is further reduced by using cubic specimens instead of conventional cylindrical ones;
e) analysis of deformation is significantly simplified by the acquisition of undistorted images.
The above technique has been employed to measure deformation and water retention of a compacted swelling clay (Wyoming bentonite) along wetting and drying paths under unconfined conditions. Tests performed on both isotropically and anysotropically compacted specimens show that there is no significant difference between the soil water retention curves of anisotropically and isotropically compacted samples (at least for the particular degree of anisotropy considered this project). It is however worth noting that the degree of anisotropy (defined as the ratio between the magnitude of principal strains) measured by image analysis is relatively small, varying from 1 to 1.05 for isotropic specimens and from 1.05 to 1.15 for anisotropic specimens.
The above measurements of strains at large scale have been complemented by a microscopy study of the same soil subjected to similar drying / wetting cycles inside an environmental scanning electron microscope (ESEM). At microscopic level, the specimen areal strain is measured by image analysis at different scales over the specimen image and for the different suction levels imposed during drying cycles inside the ESEM. Three different scales have been studied in this project, corresponding to hundreds of microns, tens of microns and microns. Measured areal strains are consistent for similar specimens, indicating a good repeatibility of the experimental technique. Measurements also reveal that the larger the measurement scale, the larger the areal strain and that there is a threshold scale (approximately 100 microns) below which strains are magnified.
The above experiments at different scales have improved fundamental understanding of soil mechanics during drying-wetting cycles and have allowed the development of a conceptual constitutive framework in which the interaction between different fabric scales can helpfully describe the phenomenological behaviour of compacted soils.
The advantages of this experimental technique are summarised as follows:
a) the disturbance caused by specimen handling and the lengthy equalisation time following the manual measurement of the sample dimensions are avoided thanks to non-contact determination of strains via image analysis;
b) the testing time is highly reduced in comparison with conventional techniques because of forced air circulation and simultaneous testing of multiple (small) specimens at different values of relative humidity;
c) high-quality images with negligible geometrical distortion are recorded thanks to the use of a single lens reflex (SLR) photographic camera - equipped with an advanced photo System type-C (APS-C) sensor - in combination with a macro lens;
d) geometrical distortion is further reduced by using cubic specimens instead of conventional cylindrical ones;
e) analysis of deformation is significantly simplified by the acquisition of undistorted images.
The above technique has been employed to measure deformation and water retention of a compacted swelling clay (Wyoming bentonite) along wetting and drying paths under unconfined conditions. Tests performed on both isotropically and anysotropically compacted specimens show that there is no significant difference between the soil water retention curves of anisotropically and isotropically compacted samples (at least for the particular degree of anisotropy considered this project). It is however worth noting that the degree of anisotropy (defined as the ratio between the magnitude of principal strains) measured by image analysis is relatively small, varying from 1 to 1.05 for isotropic specimens and from 1.05 to 1.15 for anisotropic specimens.
The above measurements of strains at large scale have been complemented by a microscopy study of the same soil subjected to similar drying / wetting cycles inside an environmental scanning electron microscope (ESEM). At microscopic level, the specimen areal strain is measured by image analysis at different scales over the specimen image and for the different suction levels imposed during drying cycles inside the ESEM. Three different scales have been studied in this project, corresponding to hundreds of microns, tens of microns and microns. Measured areal strains are consistent for similar specimens, indicating a good repeatibility of the experimental technique. Measurements also reveal that the larger the measurement scale, the larger the areal strain and that there is a threshold scale (approximately 100 microns) below which strains are magnified.
The above experiments at different scales have improved fundamental understanding of soil mechanics during drying-wetting cycles and have allowed the development of a conceptual constitutive framework in which the interaction between different fabric scales can helpfully describe the phenomenological behaviour of compacted soils.