Assessment of Liquefaction Potential Index Using Deterministic and Probabilistic Approaches

1. Introduction

In engineering design point of view, nuclear power plant structures are meant to survive all possible natural disasters, especially induced effects of earthquakes. Due to huge population and rapid urbanization, the demand of energy in India is very high. Since the energy obtained from the conventional sources is inadequate, more and more nuclear power plants are being set up to meet the increasing energy demand. The recent Fukushima Nuclear power plant disaster has created concern among people about the safety of these facilities during any natural disasters like earthquake. In this paper, the liquefaction hazard due to earthquake was assessed based on deterministic and probabilistic approaches for a nuclear power plant site situated on the east coast of South India. Liquefaction is one of the major earthquake hazards, in which the loose, cohesionless, saturated soil loses its shear strength under earthquake loading. The devastating effects of soil liquefaction were first observed during major earthquakes of Niigata (M_s = 7.5) and Alaska (M_w = 9.2) in the year 1964. In India, earthquake induced soil liquefaction was observed during Bhuj earthquake of 2001. The devastations caused by these seismic soil liquefaction have stressed the need for the assessment of liquefaction potential, especially for critical locations like nuclear power plant site.

The most widely accepted methodology for assessing liquefaction potential was proposed by Seed and Idriss (1971). In this method, the factor of safety against liquefaction was evaluated as the ratio of cyclic shear strength of soil to the cyclic stress induced in soil during earthquake. The cyclic stress induced in soil due to earthquake dependent on the peak ground acceleration (PGA), overburden pressure ratio, depth of soil layer under consideration and magnitude of the earthquake. The cyclic shear strength of the soil is evaluated either by conducting laboratory tests such as cyclic tri-axial test, cyclic simple shear test and cyclic torsional test on the undisturbed soil specimen, or from data obtained from in-situ field tests such as standard penetration test (SPT), cone penetration test (CPT), shear wave velocity test (Vs) and the Becker penetration test (BPT) etc. Difficulties associated with obtaining good undisturbed soil samples for laboratory testing and high cost of testing equipments made the evaluation of liquefaction potential based on field tests more popular.

The factor of safety against liquefaction obtained from laboratory or field tests represents the liquefaction potential of soil only at a particular depth. Hence, Iwasaki et al. (1982) has introduced liquefaction potential index (LPI), which is obtained by integrating factor of safety against liquefaction at all depths. Since it provides a unique value for a particular borehole or soil column, LPI represents liquefaction potential for that borehole in a better manner. Sonmez (2003) proposed classification of a site based on LPI values (Table 1). In the present study, LPI was estimated based on field test data using both deterministic as well as probabilistic approaches for a nuclear power plant site.