Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

An Experimental and Theoretical Study of Pile Foundations Embedded in Sand Soil

Jebur, AAJ (2018) An Experimental and Theoretical Study of Pile Foundations Embedded in Sand Soil. Doctoral thesis, Liverpool John Moores University.

2018AmeerPhD.pdf - Published Version

Download (8MB) | Preview


This study aimed to examine the load carrying capacity of model instrumented piles embedded in sand soil, and to develop and verify reliable, highly efficient predictive models to fully correlate the non-linear relationship of pile load-settlement behaviour using a new, self-tuning artificial intelligence (AI) approach. In addition, a new methodology has been developed, in which the most effective pile bearing capacity design parameters can be precisely determined. To achieve this, a series of comprehensive experimental pile load tests were carried out on precast concrete piles, steel closed-ended piles and steel open-ended piles, comprised of three slenderness ratios of 12, 17 and 25, using an innovative calibrated testing rig, designed and manufactured at Liverpool John Moores University. The model piles were tested in a large pile testing chamber at a range of different densities of sand; loose (18%), medium (51%) and dense (83%). It is worth noting that novel structural fibres were utilised and optimised for different volume fractions to enhance the mechanical performance of concrete piles. The obtained results revealed that the higher the values of the of the pile effective length, Lc (embedded length of pile), sand density, and the soil-pile angle of shearing resistance, the higher the axial load magnitudes to reach the yield limit. This can be attributed to the increase in the end bearing point and mobilised shaft resistance. In addition, the plastic mechanism occurring in the surrounding soil was identified as the leading cause for the presence of nonlinearity in the pile-load tests. Furthermore, a new enhanced self-tuning supervised Levenberg-Marquardt (LM) training algorithm, based on a MATLAB environment, was introduced and applied in this process. The proposed algorithm was trained after conducting a comprehensive statistical analysis, the key objectives being to identify and yield reliable information from the most effective input parameters, highlight the relative importance “Beta values” and the statistical significance “Sig values” of each model input variable (IV) on the model output. To assess the accuracy and the efficiency of the employed algorithm, different measuring performance indicators (MPI), suggested in the open literature, were utilised. Common statistical performance indexes, i.e., root mean square error (RMSE), Pearson’s moment correlation coefficient (p), coefficient of determination (R), and mean square error (MSE) for each model were determined. Based on the graphical and numerical comparisons between the experimental and predicted load-settlement values, the results revealed that the optimum models of the LM training algorithm fully characterised load-settlement response with remarkable agreement. Additionally, the proposed algorithm successfully outperformed the conventional approaches, demonstrating the feasibility of the current study. New design charts have been developed to calculate the individual contribution of the most significant pile bearing capacity design parameters “the earth pressure coefficient (K) and the bearing capacity factor (N )”. The improved approach takes into account the change in sand relative density, pile material type, and the pile slenderness ratios. It is therefore a significant improvement over most conventional design methods recommended in the existing design procedures, which do not consider the influence of the most significant parameters that govern the pile bearing capacity design process.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: sand soil, artificial neural network, pile foundations
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Divisions: Civil Engineering (merged with Built Env 10 Aug 20)
Date Deposited: 13 Sep 2018 10:10
Last Modified: 08 Nov 2022 14:14
DOI or ID number: 10.24377/researchonline.ljmu.ac.uk.00009211
Supervisors: Atherton, W, Al Khaddar, R and Loffill, E
URI: https://researchonline.ljmu.ac.uk/id/eprint/9211
View Item View Item