Title: Utilising recurrent neural network technique for predicting strand settlement on brittle sand and geocell

Authors: S. Jeyanthi; R. Venkatakrishnaiah; K.V.B. Raju

Addresses: Department of Civil Engineering, Bharath Institute of Higher Education and Research, Tamil Nadu, India ' Department of Civil Engineering, Bharath Institute of Higher Education and Research, Tamil Nadu, India ' Department of Civil Engineering, Bharath Institute of Higher Education and Research, Tamil Nadu, India

Abstract: In this study, geocell was used as a ground enhancement technology to enhance the tensile qualities of poor soil. Investigations have been done on the variations in settlement brought on by employing different combinations of depths of reinforcement, different layers of soil reinforcement using geocell, and varied relative densities of geocell as reinforcing material. Settlement predictions at various geocell placements in poor sand have been depicted using the recommended recurrent neural network (RNN) method. When comparing the two models, it was found that the RNN model performed better with geocell than the ANN-EHO model, the JSA model, and the MOA model. Independent variables for relative density, reinforcement depth, number of layers, and geocell height were used to provide simulation data for use in creating RNN models.

Keywords: mayfly optimisation algorithm; MOA; artificial neural network; recurrent neural network; RNN; geocell; settlement; soil reinforcement; jellyfish search algorithm; JSA.

DOI: 10.1504/IJIEI.2023.132699

International Journal of Intelligent Engineering Informatics, 2023 Vol.11 No.2, pp.122 - 137

Received: 08 Nov 2022
Accepted: 30 Mar 2023
Published online: 08 Aug 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article