Loading...

A rapid design procedure for tied-back soil walls using multivariate adaptive regression splines (MARS) method

Maleki, J ; Sharif University of Technology | 2023

0 Viewed
  1. Type of Document: Article
  2. DOI: 10.1007/s10706-022-02351-y
  3. Publisher: Springer Science and Business Media Deutschland GmbH , 2023
  4. Abstract:
  5. Deep excavations are widespread geotechnical structures in urban areas. Vertical deep excavations may cause irreparable harms to the adjacent structures due to excessive deformations. Therefore, obtaining accurate soil parameters, rigorous design of wall’s structural components, and monitored construction are essential elements of any successful project of this kind. Soil-nail walls are a type of tied-back walls that are widely used due to their simple execution and low cost. In order to facilitate the design procedure of Soil-nail walls for deep excavations and reduce the number of numerical modelings, Multivariate Adaptive Regression Splines (MARS) method has been employed. The developed MARS model is based on a large database that has been provided using FEM analyses of Soil-nail walls behavior with various depths, surcharges, soil types, and nailing patterns. By using the proposed design procedure, one can come up with a preliminary design of Soil-nail wall for deep excavation close to the optimal design based on FHWA (2015) guidelines which can be further improved regarding specific issues of the site. Also, by using the proposed method, one can have a fairly accurate estimation of the maximum deflection of the wall to check whether or not the design can satisfy the safety requirements for the adjacent structures. © 2022, The Author(s), under exclusive licence to Springer Nature Switzerland AG
  6. Keywords:
  7. Deep excavation ; MARS model ; Maximum horizontal deformation ; New design parameter ; Optimal design ; Rapid design ; Soil-nail wall
  8. Source: Geotechnical and Geological Engineering ; Volume 41, Issue 2 , 2023 , Pages 1521-1535 ; 09603182 (ISSN)
  9. URL: https://link.springer.com/article/10.1007/s10706-022-02351-y