Repeated small shaking events due to earthquakes significantly enhance liquefaction resistance of soils. Analyses of liquefaction case histories show that aged soils in seismically active zones tend to be less vulnerable to liquefaction despite having similar index parameters—such as standard penetration test N-values and shear wave velocities—as young soils. Significant efforts have been devoted to better understand the effects of the cyclic pre-shearing on liquefaction resistance and it was found that this effect depends on the number of cycles and cyclic stress ratio. However, none of these parameters quantify the improvement of liquefaction resistance due to pre-shaking. This study investigates the pre-shearing effects on liquefaction resistance through laboratory tests and centrifuge tests. An attempt was made to explain the effects quantitatively with a single index parameter of the volumetric strain caused by pre-shearing. It was confirmed from triaxial tests that the liquefaction resistance of pre-sheared sand uniquely increased with increasing volumetric strain regardless of the cyclic shear stress ratio and the number of cycles during the pre-shearing. To examine the pre-shaking effects on the liquefaction strength of sand under a level ground condition, centrifuge tests were conducted in this study. Sand models were subjected to small shaking events repeatedly, which were weak enough not to cause liquefaction. It was observed that changes in the index parameters of the models, including soil density (volumetric strain), shear wave velocity, and horizontal earth pressure during the pre-shaking events were very small. At the end of the test, the sand was subjected to a strong shaking event because models that had gone through pre-shaking need larger shaking acceleration to liquefy. Liquefaction resistance was derived from acceleration records with the aid of the cumulative damage theory. The relationship between liquefaction resistance ratio and volumetric strain that occurred in the pre-shaking events coincides with the relationship obtained from the triaxial tests. After the extensive liquefaction event, all index parameters except soil density—K0, Vs, liquefaction resistance—tended to return to their original values (before the pre-shaking).

Introduction

The resistance to liquefaction of sandy soils that have been resting for many years is greater than that of recently deposited soils. This aging effect on liquefaction resistance may be explained by two mechanisms. One is the improved interlocking of sand grains developed after deposition, which is associated with their extended time under static pressure and being subjected to repeated earthquake shakings. The other mechanism is the long period of sustained static load [17] that is probably associated with such chemical reactions as dissolution of minerals and precipitation at soil grain surface, which develop bonding between soil particles. The focus of this paper is on the first mechanism. It was pointed out that repeated small shakings due to earthquakes significantly enhance soil liquefaction resistance. Analyses of liquefaction case histories showed that older soils in seismically active zones tend to be less vulnerable to liquefaction although their index parameters—such as standard penetration test N-values and shear wave velocities—were very similar [4,11]. Moreover, these facts are supported by many laboratory tests [7,9,12,19,20]. In view of these, significant efforts in laboratory testing were devoted to better understand the effects of cyclic pre-shearing on liquefaction resistance. The test results consistently indicated that the liquefaction resistance increased with the number of cycles and cyclic stress ratio provided that the shear strain during pre-shearing was small. However, none of these parameters quantify the improvement of the liquefaction resistance due to pre-shaking.