01628129

Component

The paper presents an application of three methods: regression analysis, artificial neural networks (ANNs) and support vector machines (SVMs), for the estimation of the compaction parameters: maximum dry density (MDD) and optimum moisture content (OMC) from index properties of the soils: liquid limit (LL), plastic limit (LP), plasticity index (PI), grain-size distribution and specific gravity (Gs). The data collected in the course of laboratory testing was used for the estimation of soil compaction parameters. The samples belong to various clay types, and were obtained from cores from four earth-fill dams: Rovni, Selova, Prvonek and Barje, located in Serbia and served as control samples during soil compaction. The developed models can be used to estimate the compaction parameters: (i) in the preliminary stages of the project development, and (ii) in the course of the preliminary assessment of the suitability of a material from borrow pits for use in earth-fill structures. This analysis also shows the comparison between the three methods in terms of applicability and goodness of fit.

This paper was sponsored by TRB committee AFS20 Standing Committee on Geotechnical Instrumentation and Modeling.

01618707

17-02100

English

13p

2017

Transportation Research Board 96th Annual Meeting

Digital/other

Figures; References; Tables

Compacted clays; Compaction; Dams; Dry density; Estimating; Laboratory tests; Machine learning; Moisture content; Neural networks; Regression analysis; Soils; Vector analysis

Serbia

Geotechnology; Transportation (General)

Transportation Research Board Annual Meeting 2017 Paper #17-02100

TRIS, TRB, ATRI

Dec 8 2016 10:45AM