Investigation of dynamic behavior of back to back reinforced soil retaining walls with finite element and artificial neural network

Back-to-back Mechanically Stabilized Earth (MSE) wall are commonly used for bridge approach embankments. By the means of not only aesthetically pleasing appearance but also satisfactory performance under seismic loading reinforced soil retaining structures are becoming widely used in Turkey.In this study, a parametric study of seismic response analysis of reinforced soil retaining structures was performed using a finite element analysis with commercial Finite Element software, Plaxis. The aim of the study is to determine the influence of reinforcement length, reinforcement spacing, wall height and facing type on seismic-induced permanent displacements. Permanent displacements under earthquake loading conditions associated with different L/H ratios and reinforcement spacing for 5m, 6m, 7m, 8m, and 9m height walls are investigated. In order to investigate dynamic behavior of the walls harmonic motions with 5 seconds duration have been applied. The harmonic motion had three different levels of Peak Ground Accelerations, namely 0.2g, 0.4g and 0.6g. The frequency of the harmonic input acceleration record is 3 Hz.Artificial Neural Network conducted in this study was applied for the first time in literature to estimate the deformations of retaining walls under dynamic loads. Although developing an analytical or empirical model is feasible in some simplified situations, most manufacturing processes are complex, and therefore, models that are less general, more practical, and less expensive than the analytical models are of interest. ANN has been applied successfully in many non-linear geotechnical engineering problems in order to make reliable predictions and to check whether the range of classical design results are within reasonable outcomes.