Laboratory testing is an essential addition to field investigations. Due to lack of finances, poor planning and absence of specific permafrost knowledge this is hardly ever carried out for mountain permafrost projects, in particular in the Alps, except for simple soil classifications. It is, however, crucial when determining geothermal and geotechnical soil properties because these parameters can be measured in a controlled environment and extreme boundary conditions can be applied. It is moreover important to test the soil at the boundary conditions that are to be expected in the field. Laboratory testing includes:
o Soil Classification: Standard soil classification includes determination of ice content. Additional information, such as unfrozen water contents are recommended for high risk projects.
o Thermal needle probe: The thermal conductivity of a soil can be measured with a thermal needle (ASTM D5334-05, 2005). This parameter is needed for thermal modelling. Pushing the needle into a coarse frozen sample may be difficult to do.
o Thermal plate: Similar to the thermal needle, but no needle is pushed into the soil and the thermal conductivity can be measured from the sample surface.
o Uniaxial compression test: Uniaxial compression tests are the simplest way to determine creep or strength parameters for a frozen soil (ASTM, 2001; ASTM, 2006). However, the strengthening effect of the confinement can not be determined and the strain rates must be chosen carefully. Strain rates of 1%/min, as recommended by ASTM, may result in brittle behaviour that overestimates the long term strength.
o Triaxial compression test: Even though triaxial 'constant strain rate' and 'constant stress' (creep) tests are difficult and time consuming and are highly dependent on sample size and quality, currently they provide the best insight into deformation and failure mechanisms.
o Direct shear test: In order to study distinct shear interfaces, such as permafrost tables or rock joints, direct shear tests are highly recommended. These tests can be carried out in the laboratory (Yasufuku et al., 2003) or in the field (Springman et al., 2003).
o Centrifuge modelling: Centrifuge modelling (e.g., Harris et al., 2008; Harris et al., 2001b; Kern-Luetschg et al., 2008) can be a useful investigation tool for studying certain processes. Soil properties, however, can not be directly measured and scaling laws must be considered, in particular when modelling coarse soils.
The higher the risk (Consequence - Sensitivity matrix), the more knowledge about the soil's properties is required. A thorough laboratory testing programme that includes triaxial or direction shear testing is strongly advised. Depending on the project, even more complex studies that involve centrifugal modelling may be considered because they offer an opportunity to study failure mechanisms on a larger scale.
Was this article helpful?