Effect of in-situ sample quality on undrained cyclic strength and liquefaction assessment

This study investigates the effects of quality of in-situ sandy samples, which were taken by a conventional tube sampling (TS) method, on liquefaction assessment. The sandy samples were retrieved from a loose fill layer and medium-dense Holocene deposit in young reclaimed land in Urayasu City where significant liquefaction occurred during the 2011 off the Pacific Coast of Tohoku Earthquake. First, a series of small strain measurements and undrained cyclic triaxial tests was conducted on the TS and reconstituted samples. Second, in order to investigate an applicability of the tested samples for the liquefaction assessment, an effective stress analysis was performed by use of soil parameters obtained from the examined samples. Input earthquake motions of the analysis are the 2011 off the Pacific Coast of Tohoku Earthquake and the 1987 Chiba Toho-oki Earthquake, the first and the second largest earthquakes the investigation site has experienced. The result of a sample quality assessment with the small strain shear moduli indicates that the TS sample retrieved from the Holocene deposit remains of high quality. In addition, for the reconstituted sample of the fill layer, since the small strain shear moduli of the sample are found to be nearly equal to those obtained at the field, it can be considered that the in-situ soil structure condition is reproduced in the reconstituted sample. According to an official report, the liquefaction resistance of the TS sample of the loose fill layer is RL20=0.38, which is much larger than that of medium-dense Holocene deposit (RL20=0.29) measured in this study. This is because the fill layer is so loose that the TS sample could have been strengthened due to disturbance during the sampling. On the other hand, the test result for the reconstituted sample of fill layer shows an appropriate level of liquefaction resistance (RL20=0.14). The results of the effective stress analysis using the soil parameters obtained from the TS sample of Holocene deposit and the reconstituted sample of fill layer show the most probable behaviors of the ground at the investigation site for both the 2011 off the Pacific Coast of Tohoku Earthquake and the 1987 Chiba Toho-oki Earthquake, which are consistent with the evidence reported after the earthquakes. The results indicate that the sample quality assessment by using a small strain shear moduli is applicable for the rational liquefaction assessment, and it would be possible to employ the reconstituted sample, which the small strain shear moduli are the same with the in-situ value, in substitution for a high quality sample.