抄録
The effects of sample preparation methods on the cyclic undrained stress-strain behavior of sands were investigated by means of triaxial and torsional shear tests using two kinds of clean sands and one kind of sand including fines, with a wide variation in the sample density. Four different sample preparation methods were adopted ; air-pluviation, wet-tamping, wet-vibration, and water-vibration. The differences in the effects of sample preparation method were found to be significant in both testing methods. However, specific ways in which the sample preparation methods affected the results were not consistent between the triaxial and torsional shear tests. This indicates that the process of estimating cyclic undrained simple shear strengths from triaxial strengths is not as simple as has previously been considered.It was found that in torsional shear tests the shapes of strength curves in the planes of the cyclic stress ratio and the logarithm of the number of cycles Nc are similar for different sample preparation methods, in the sense that the different curves collapse into a single curve when these curves are moved appropriately along the log Nc-axis. It was also found that for the same sample density the effects of the sample preparation method on the cyclic undrained stress-strain behavior decrease with the increase in strain. A new density index is proposed, from which the cyclic undrained behavior can be estimated for different sample preparation methods and for different kinds of sands until the shear strain becomes around 3% in double amplitude.
