Evaluation of drainage ability of granular subbase through large-scale model pavement studies and machine learning models

The detrimental effects of heavy wheel loads on pavements with saturated subbase materials are deemed to be deciding factors on the serviceability of pavements. The major damages owing to this are, immediate premature failures and reduced strength, culminating in severe deformation and decreased lon...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Innovative infrastructure solutions : the official journal of the Soil-Structure Interaction Group in Egypt (SSIGE) 2024-03, Vol.9 (3), Article 82
Hauptverfasser: Kavitha, Govindaswamy, Kunjar, Vaishakh, Gowda, Sachin, Jayaram, Mulukunte Anantharamaiah, Veeraragavan, Amirthalingam
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The detrimental effects of heavy wheel loads on pavements with saturated subbase materials are deemed to be deciding factors on the serviceability of pavements. The major damages owing to this are, immediate premature failures and reduced strength, culminating in severe deformation and decreased longevity of the pavements. This paper presents a three-phase approach to evaluate the drainage ability of varying Granular Subbase mixes used as subbase layers. In the first phase, explorative laboratory and large-scale model pavement studies on the hydraulic conductivity ( k ) of different GSB mixes of varying gradations recommended by the Ministry of Road Transport and Highways (MoRT&H), Indian guidelines were performed. The results of the experiments were used to compare the horizontal hydraulic conductivity of these mixes at varying hydraulic gradients with and without geotextile lining. In the second phase, rainfall simulation studies replicating the field conditions were conducted to determine the optimal gradation for effective drainage. The findings revealed that open grades having higher effective size ( D 10 ) values of 2.24 mm and 1.70 mm showed higher hydraulic conductivity of compared to the close grades. It was also found that the use of geotextile lining reduced the k -value of all the tested grades due to the clogging effect. In the third phase, the data generated were used to develop an optimal machine learning model. Principal component analysis revealed D 10 , porosity, and particles finer than 0.075 mm ( P 0.075 ) as significant parameters influencing the hydraulic conductivity. Among artificial neural network (ANN) and regression models, the ANN model provided accurate functional mapping between horizontal hydraulic conductivity values and the influencing factors with a high R 2 value of 0.98. As an outcome of this research a chart was developed to determine the thickness of the granular subbase layer for the different rainfall conditions.
ISSN:2364-4176
2364-4184
DOI:10.1007/s41062-024-01391-y