PEER Workshop on Liquefaction Susceptibility

Abstract

Seismic ground failure potential from liquefaction is generally undertaken in three steps. First, a susceptibility evaluation determines if the soil in a particular layer is in a condition where liquefaction triggering could potentially occur. This is followed by a triggering evaluation to estimate the likelihood of triggering given anticipated seismic demands, environmental conditions pertaining to the soil layer (e.g., its depth relative to the ground water table), and the soil state. For soils where triggering can be anticipated, the final step involves assessments of the potential for ground failure and its impact on infrastructure systems. This workshop was dedicated to the first of these steps, which often plays a critical role in delineating risk for soil deposits with high fines contents and clay-silt-sand mixtures of negligible to moderate plasticity. The workshop was hosted at Oregon State University on September 8-9, 2022 and was attended by 49 participants from the research, practice, and regulatory communities. Through pre-workshop polls, extended abstracts, workshop presentations, and workshop breakout discussions, it was demonstrated that leaders in the liquefaction community do not share a common understanding of the term “susceptibility” as applied to liquefaction problems. The primary distinction between alternate views concerns whether environmental conditions and soil state provide relevant information for a susceptibility evaluation, or if susceptibility is a material characteristic. For example, a clean, dry, dense sand in a region of low seismicity is very unlikely to experience triggering of liquefaction and would be considered not susceptible by adherents of a definition that considers environmental conditions and state. The alternative, and recommended, definition focusing on material susceptibility would consider the material as susceptible and would defer consideration of saturation, state, and loading effects to a separate triggering analysis. This material susceptibility definition has the advantage of maintaining a high degree of independence between the parameters considered in the susceptibility and triggering phases of the ground failure analysis. There exist differences between current methods for assessing material susceptibility – the databases include varying amount of test data, the materials considered are distinct (from different regions) and have been tested using different procedures, and the models can be interpreted as providingdifferent outcomes in some cases. The workshop reached a clear consensus that new procedures are needed that are developed using a new research approach. The recommended approach involves assembling a database of information from sites for which in situ test data are available (borings with samples, CPTs), cyclic test data are available from high-quality specimens, and a range of index tests are available for important layers. It is not necessary that the sites have experienced earthquake shaking for which field performance is known, although such information is of interest where available. A considerable amount of data of this type are available from prior research studies and detailed geotechnical investigations for project sites by leading geotechnical consultants. Once assembled and made available, this data would allow for the development of models to predict the probability of material susceptibility given various independent variables (e.g., in-situ tests indices, laboratory index parameters) and the epistemic uncertainty of the predictions. Such studies should be conducted in an open, transparent manner utilizing a shared database, which is a hallmark of the Next Generation Liquefaction (NGL) project.