Evaluation of Liquefaction-Induced Lateral Spreading by a Neural Network (NN) Model

Abstract eng:
Liquefaction-induced lateral spreading during an earthquake ground motion (EQGM) is considered to be one of the major soil failure types. The observations from the previous earthquakes show that the magnitude of lateral spreading can reach to the level of a few meters that can severely damage the surrounding area most of the times. Since liquefaction is a complex soil failure problem that involves soil and earthquake parameters, liquefaction-induced soil deformations should be investigated by nonlinear methods. In this study, a NN model has been developed to investigate the liquefaction-induced lateral spreading during an earthquake ground motion. The NN model used in this study was constructed using 8 input parameters (magnitude of EQGM, σ0 (kPa), σ0' (kPa), N (SPT), a/g, τort/σ0, F (%), D50) and 1 output parameter (lateral spreading). A relationship is established between the lateral spreading and the soil and earthquake ground motion parameters by developing and testing a multi-layered feed-forward NN model trained with the back-propagation algorithm. A total of 175 cases were used for constructing the NN model. Results indicates that N (SPT) is the most effective parameter; whereas a/g is the least effective parameter on liquefaction-induced lateral spreading. The approach adapted in this study was shown to be capable of providing the best accurate estimates of liquefaction-induced lateral spreading during an earthquake ground motion.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket