We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
An extreme learning neural network approach for seismic bearing capacity estimation of planar caissons in nonhomogeneous clays.
- Authors
Lai, Van Qui; Chauhan, Vinay Bhushan; Keawsawasvong, Suraparb; Sangjinda, Kongtawan; Chavda, Jitesh T.; Mase, Lindung Zalbuin
- Abstract
This study employs a two-dimensional plane strain finite element limit analysis method to evaluate the seismic bearing capacity of a planar caisson in anisotropic and non-homogeneous clay. The anisotropic behavior of the clay is simulated using the Anisotropic Undrained Shear (AUS) failure criterion in the finite element limit analysis (FELA). A rigid caisson has a depth (L) and a width (B). A comprehensive parametric analysis is executed to evaluate the non-dimensional seismic bearing capacity factor (Nce) in terms of the adhesion factor (α), anisotropic strength ratio (re), horizontal seismic coefficient (kh), depth to diameter ratio (L/D), and shear strength gradient ratio (ρB/suc0). The relationship between these parameters to the seismic bearing capacity factor is investigated, and the influence of these parameters on the potential failure mechanisms is discussed in detail. Moreover, an equation for predicting the seismic bearing capacity factor is developed through a machine learning regression approach called the Artificial Neural Network (ANN) model, which practitioners can extensively employ in the field. These correlation functions fit well with those obtained from FELA, with a value of R2 = 99.43%.
- Subjects
SEISMIC networks; MACHINE learning; CAISSONS; FINITE element method; CLAY; SHEAR strength
- Publication
Earth Science Informatics, 2024, Vol 17, Issue 1, p251
- ISSN
1865-0473
- Publication type
Article
- DOI
10.1007/s12145-023-01175-5