|
Eniang, I. S.
Department of Civil Engineering, Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria
Arinze, E. E.
Department of Civil Engineering, Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria
ABSTRACT
This study explored the application of the Artificial
Neural Network (ANN) technique to develop a back propagation model for the
prediction of volume of fill materials as highway embankments. A total of sixty
(60) lateritic and clay soil samples were collected through disturbed sampling
from a borrow pit at an embankment construction site in Akwa Ibom State, at
depths of 1m to 2m. Various tests, including moisture content, Atterberg
limits, compaction, and California Bearing Ratio (CBR) tests, were performed on
both the soil samples and the reinforced rigid pavement. The ANN method was
employed to create a model for the prediction of volume of fill material.
Results from the Atterberg limits test indicated that most of the soil samples
were suitable for use as fill material, with liquid limits (LL) and plasticity
index (PI) values below 35% and 12%, respectively. Specific gravity values were
within specification, further confirming the soil's suitability for highway
embankment fill. The CBR test results showed a range of values of no less than
12%, with an average of 12.68%, and sample 49 exhibited the highest value
(15.5%) after 48 hours of soaking. The ANN model demonstrated a notable
improvement over the MLR model, achieving an R² value of 0.661.
Keywords: ANN, construction, embankment, highway pavement
https://doi.org/10.33922/j.ujet_v11i1_5
|
View: 41 | Download: 6
Published
Monday, February 03, 2025
Issue
Vol. 11 No. 1, June 2025
Article Section
GENERAL
The contents of the articles are the sole opinion of the author(s) and not of UJET.
|
