Performance-based assessment of liquefaction-induced free-field ground settlement

抄録

The accumulation of liquefaction-induced volumetric strains leads to ground settlement. Volumetric strain is estimated using empirical models based on laboratory data, and settlement is computed as the integration of volumetric strain with depth in the liquefied materials and calibrated using field data. Currently, the assessment of liquefaction-induced free-field ground settlement (S_v) generally follows a deterministic or a pseudo-probabilistic approach. In both cases, the assessment of the ground motion intensity measure (IM) is decoupled from the computation of S_v. For example, a hazard curve for the IM is developed through a probabilistic seismic hazard assessment that accounts for all relevant earthquake scenarios in a pseudo-probabilistic approach. A design hazard level for the IM is defined, and S_v is computed based on this design IM. A key assumption in this approach is that the hazard level of the IM is consistent with the hazard level for S_v. However, this assumption is not always valid. A performance-based approach for the assessment of S_v is developed in which the hazard evaluation for the IM is explicitly incorporated in the assessment of S_v by combining the hazard curve for the IM with the probability of exceeding different S_v levels. Hence, the sources of variability contributing to the IM are incorporated in the assessment of S_v. The variability in the inputs to the empirical model for S_v can also be included through a logic-tree approach. As a result, the hazard curve for S_v is developed, which directly links different hazard levels with their corresponding values of ground settlement. Conventional approaches used currently do not always produce values of settlement that are compatible with design hazard levels.