Risk Assessment And Hazard Prediction Of Seismic Liquefaction Using Response Surface Method

Liquefaction is one of the most damaging functions of earthquakes in saturated sandy soil.Therefore,clearly advancing assessment of liquefaction risk and hazard is one of the key points for geotechnical profession.There are three main necessary aspects for geotechnical engineers to consider in order to performing this assessment:1)liquefaction susceptibility 2)evaluating sufficient dynamic loads for triggering liquefaction 3)effect of liquefaction.Evaluation of first two aspects has been inspected in analyzing potential of liquefaction that states the liquefaction will occur or no.Furthermore,evaluation of the damaging effects of liquefaction,mostly lateral displacement,is the last aspect to perform hazard analysis.Among all approaches have been introduced by researchers,empirical and semi-empirical models using case histories or laboratory test results are the most common and easy to use by engineers and researchers.The intelligent algorithm has been applied to discover correlation between parameters and the hazard factors.Nevertheless,it is still lacking the validating phase to prevent the overtraining of the models.While,the complex influence of the fine content(FC)on generation of pore water pressure has been demonstrated,there is no consideration to this parameter in available models.All models contain this parameter without any limitations on its value.Moreover,it has been demonstrated that standardized cumulative absolute velocity(CAV5)provides the most sufficient and efficient intensity motions to rising water pressure and liquefaction.However,the momentum magnitude(Mw)and horizontal peak ground acceleration(PGA)are traditionally correlated to liquefaction hazard assessment in available models.Further,due to natural variability and insufficient knowledge in soil properties,geometry conditions and earthquake loads,geotechnical problems,particularly high non-linear conditions such as liquefaction phenomenon,encounter with uncertainties.To govern with this uncertainties and quantified them,the monte calro simulation(MCS)using artificial neural network(ANN)-based is introduced in this study to conduct parametric sensitivity analysis to study the influence of these uncertainties on the liquefaction hazard.The response surface method(RSM)is applied as a novel method to inspect its applicability and ability with different well known design of experiment(DOE),as far as possible,to compare their effects on the liquefaction hazard.Through adding CAV5 to the dataset.significant aspects of earthquake such as:near fault zone.frequency content of earthquake's motion,causative fault type and duration of the earthquake is considered to derive a capable model to evaluate near fault zones hazard.