LEAP (Liquefaction Experiment and Analysis Project)

Introduction

Resource: M. Manzari, B. Kutter and M. Zeghal (2017), “Validation of Constitutive and Numerical Modeling Techniques for Soil Liquefaction Analysis using LEAP.”
Over the last three decades, the geotechnical earthquake engineering field has witnessed tremendous advances in experimental and computational simulation capabilities of soil liquefaction. Soil element laboratory tests, in-situ tests, and centrifuge experiments are increasingly more sophisticated and reliable and provide a wealth of experimental data that is continuously improving our understanding of the response of geostructures to earthquake excitations. Numerous constitutive models were developed and many are now well-established and capable of predicting the intricate response of saturated granular soils under extreme loading conditions, such as the transition from a contractive to a dilative response. Several open-sources and commercial computational codes are available and provide fully-coupled (two-phase) effective-stress nonlinear analysis of geosystems involving liquefiable soils. These numerical tools offer a valuable supplement to experimentation that enables a significant reduction in needed testing programs of soil sample and small scale (physical) model tests, leading to substantial time and financial savings. Practitioners and academics acknowledge the value of numerical simulations, but often raise the credibility issue. There is little trust in numerical predictions without verification and validation (V&V), replication (consisting of repeatability and reproducibility or R&R of tests), and uncertainty quantification (UQ).

 

The Liquefaction Experiments and Analysis Projects (LEAP) is an international effort to produce a set of high quality test data and then use it in a validation exercise of existing computational models and simulation procedures for soil liquefaction analysis.

Planning Project

Resource: M. Manzari, B. Kutter and M. Zeghal (2017), “Validation of Constitutive and Numerical Modeling Techniques for Soil Liquefaction Analysis using LEAP.”

A planning LEAP (or P-LEAP) was conducted in 2015. A series of system identification analyses had shown that a level site or sloping ground tested in laminar box is associated with rather complex boundary conditions. These conditions make the task of calibration and validation of numerical modeling techniques quite difficult. Also, the laminar boxes used by different centrifuge facilities are often substantially different from each other. We concluded that a test in a rigid container provides analogous data at different centrifuge facilities, is simpler to simulate in a numerical model, and, hence, could provide more reliable data for validation purposes. A relatively simple centrifuge experiment of a sloping deposit tested in rigid container was designed to address the validation of numerical tools in predicting earthquake induced lateral spreading of a liquefiable soil. In coordination with our international collaborators, five centrifuge tests were carried out at the facilities of UC Davis (UCD), Rensselaer (RPI), Cambridge University (CU), Kyoto University (KU), National Central University of Taiwan (NCU), and Zhejiang University (ZU). Class A prediction exercises were performed by five groups of researchers from US and Japan to assess the capabilities of a few prominent constitutive models and numerical modeling techniques. An international workshop (LEAP-GWU-2015) was organized to get feedback from the geotechnical engineering community and establish protocols and procedures for verification and validation of numerical simulation tools of liquefaction phenomena. Building on the results and experiences gained from the P-LEAP, a full Liquefaction Experiment and Analysis Project (LEAP) was proposed to formally address the issue of replication of experiments and numerical predictions and investigate a number of soil liquefaction phenomena that still constitute a challenge in numerical modeling and simulations.