Buy article PDF
The purchased file will be sent to you
via email after the payment is completed.
US$ 35
|
Geomechanics and Engineering Volume 34, Number 5, September10 2023 , pages 561-575 DOI: https://doi.org/10.12989/gae.2023.34.5.561 |
|
|
|
A hybrid approach to predict the bearing capacity of a square footing on a sand layer overlying clay |
||
Erdal Uncuoglu, Levent Latifoglu and Zulkuf Kaya
|
||
| Abstract | ||
| This study investigates to provide a fast solution to the problem of bearing capacity in layered soils with easily obtainable parameters that does not require the use of any charts or calculations of different parameters. Therefore, a hybrid approach including both the finite element (FE) method and machine learning technique have been applied. Firstly, a FE model has been generated which is validated by the results of in-situ loading tests. Then, a total of 192 three-dimensional FE analyses have been performed. A data set has been created utilizing the soil properties, footing sizes, layered conditions used in the FE analyses and the ultimate bearing capacity values obtained from the FE analyses to be used in multigene genetic programming (MGGP). Problem has been modeled with five input and one output parameter to propose a bearing capacity formula. Ultimate bearing capacity values estimated from the proposed formula using data set consisting of 20 data independent of total data set used in MGGP modelling have been compared to the bearing capacities calculated with semi-empirical methods. It was observed that the MGGP method yielded successful results for the problem considered. The proposed formula provides reasonable predictions and efficient enough to be used in practice. | ||
| Key Words | ||
| finite element method; layered soils; loading test; multigene genetic programming; ultimate bearing capacity | ||
| Address | ||
| Erdal Uncuoglu, Levent Latifoglu and Zulkuf Kaya: Department of Civil Engineering, Erciyes University, Koşk District, Ahmet El Biruni Stress, 38030, Melikgazi, Kayseri, Turkey | ||