Deformation characteristics of soil under sequential dynamic torsional and cyclic axial loading

Abstract

Earthquake-induced structural damages caused due to site effects, soil liquefaction failure, and associated excessive settlement are well-known worldwide. The primary controlling factor responsible for the superstructure’s response under such dynamic (earthquake) loading conditions depends on the behaviour of local soil properties underneath it. These dynamic properties are usually measured in the laboratory by performing the shear deformation characteristics test employing a resonant column and cyclic triaxial test. The small strain (10-05 to 0.01 %) test is mainly performed using a resonant column, and the medium to large strain (> 0.01 % to 1 %) test is performed using a cyclic triaxial test. Thus, measurement of these dynamic properties often requires a minimum of two or multiple specimens, which involves immense effort and expense in facilitating the low and high strain loading separately. In the field, however, the characteristics of seismic waves propagating from a wave source to a site are a complex combination of different types of waves, including compressional and shear waves. Therefore, in the laboratory, it is crucial to test soils under the combined loadings of dynamic torsional and cyclic axial loading (using a single specimen) to mimic the field condition as practically as possible. The main objective of this research is to study the deformation characteristics of soils at low to high strain levels while a single specimen is subjected to combined dynamic torsional and cyclic axial (sequential) loading. For this purpose, an operational resonant column and cyclic triaxial apparatus were modified with advanced instrumentations, and dynamic tests were performed on one hundred twenty-one test specimens. Reconstituted clean sands, including a United Kingdom’s sand and in-situ soil samples (Indo-Gangetic plain region, West and South India Region, North-Eastern Region soils, and liquified soil deposits) were selected for the study. The influence of different parameters on the dynamic properties such as the confining pressure, relative density (void ratio), particle shape, number of loading cycles, soil type, and previous loading history (or pre-straining effect) was studied explicitly. After extensive studies of dynamic properties under dry conditions, dynamic behaviour under full saturated conditions, i.e., liquefaction and re-liquefication of silty sand, riverbed sand, and in-situ soils, were explored. The study brings out the following significant contributions as a part of the research work: 1) A low-cost novel Foldscope approach was introduced for quantifying the particle shape and imaging of the soil particles 2) Established appropriate methods and number of loading waveforms required for the reliable damping estimation using Resonant Column tests (at low strain levels) 3) A modified method was proposed for estimating the damping for non-symmetric hysteresis loop undergoing cyclic loading (at high strain levels) 4) Established the effects of confining pressure, relative density (void ratio), particle shape, number of loading cycles, soil type, and previous loading history (or pre-straining effect) on dynamic properties of Sand and Silty Sand 5) Shear modulus and damping model was developed for sands and natural soils with fines 6) Shear modulus and damping values were obtained for strain greater than 1 %, and a mechanism for unique behaviour of damping was established 7) Investigation of liquefaction and re-liquefaction potential of the liquefied site was made through field and laboratory experiments. Finally, a liquefaction strength curve was developed One of the key aspects of this study is that the design curves developed in this study can be incorporated into practice, such as in carrying out any dynamic analysis.