Assessment of deflection of pile implanted on slope by artificial neural network

• Published in: Neural computing & applications Vol. 33; no. 4; pp. 1091 - 1101
• Main Authors: Goudjil, Kamel, Arabet, Leila
• Format: Journal Article
• Language: English
• Published: London Springer London 01.02.2021; Springer Nature B.V
• Subjects: Artificial Intelligence; Artificial neural networks; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Deflection; Finite element method; Image Processing and Computer Vision; Mathematical models; Neural networks; Numerical models; Original Article; Probability and Statistics in Computer Science; 3D Plaxis; Mechanical and geometrical parameters of the soil; Horizontal displacement; Artificial neural networks; Pile; Slope
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-020-04985-6

The influence of the proximity of slope on the laterally loaded piles has been the subject of several researches. The main purpose of this study is developing a neural model able to predict the deflection of laterally loaded piles placed near a slope. To achieve this goal, we divided this work into two parts. In the first part, we have carried out a numerical modeling of an experimental model (Bouafia and Bouguerra in Fr Geotech J 75(2):47–56, 1996) using the three-dimensional finite element method. The relative error between the results of experimental and numerical model varies from 0.055 to 0.422, which indicates a good agreement. In the second part, after having validated the numerical model, we have created a database relating the deflection of pile to their contributory parameters. Artificial neural networks used this database in order to predict the pile deflection. The results obtained are very satisfactory with very acceptable errors ( R 2  = 0.9647, RMSE = 0.0133 and MAE = 0.0066).