Development of Neural Network-Based Asphalt Mix Design Parameters Prediction Tool

Basit Öğe Kaydı
dc.authoridKAYA, ORHAN/0000-0001-6072-3882
dc.contributor.authorKaya, Orhan
dc.date.accessioned2025-01-06T17:36:29Z
dc.date.available2025-01-06T17:36:29Z
dc.date.issued2023
dc.description.abstractMix design of asphalt concrete is carried out in order to produce a mixture of asphalt bitumen and aggregate that satisfies both minimum design requirements and cost-effectiveness. Marshall mix design, similar to other mix design methods, aims to determine an optimum mix for a given design criteria, having an optimum bitumen content, and has still been one of the most used asphalt mix design method in the world. Marshall mix design procedure requires significant amount of time and skilled workmanship for lengthy laboratory experiments. Time required for the mix design may negatively affect construction schedule especially during the peak construction season. In Marshall mix design procedure, optimum bitumen content is determined based on Marshall stability and flow test results, and other volumetric properties of the mixes. Prediction models that quickly predict all required Marshall mix design parameters based on very few numbers of input parameters requiring significantly less amount of time to obtain compared to the time needed to obtain all Marshall mix design parameters through lengthy laboratory experiments could be quite useful. In this study, artificial neural network-based prediction models based on the data of 200 asphalt mixes designed for asphalt wearing course, obtained from 5th district of the General Directorate of Highways of Turkey, were developed to predict all required Marshall mix design parameters. As part of the paper, a Microsoft Excel Macro-based tool, to be potentially used by 5th district of the General Directorate of Highways of Turkey, was also developed that makes prediction of all required mix design parameters for any given mix in seconds using the developed models and very few numbers of user-defined input parameters requiring significantly less amount time to obtain so that optimum bitumen content could be quickly determined. This will significantly reduce the amount of time and other resources required to determine optimum bitumen content.
dc.identifier.doi10.1007/s13369-022-07579-7
dc.identifier.endpage12804
dc.identifier.issn2193-567X
dc.identifier.issn2191-4281
dc.identifier.issue10
dc.identifier.scopus2-s2.0-85145095029
dc.identifier.scopusqualityQ1
dc.identifier.startpage12793
dc.identifier.urihttps://doi.org/10.1007/s13369-022-07579-7
dc.identifier.urihttps://hdl.handle.net/20.500.14669/1901
dc.identifier.volume48
dc.identifier.wosWOS:000905892500001
dc.identifier.wosqualityQ2
dc.indekslendigikaynakWeb of Science
dc.indekslendigikaynakScopus
dc.language.isoen
dc.publisherSpringer Heidelberg
dc.relation.ispartofArabian Journal For Science and Engineering
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanı
dc.rightsinfo:eu-repo/semantics/closedAccess
dc.snmzKA_20241211
dc.subjectAsphalt mix design
dc.subjectMarshall mix design
dc.subjectArtificial neural network
dc.titleDevelopment of Neural Network-Based Asphalt Mix Design Parameters Prediction Tool
dc.typeArticle

Dosyalar

Koleksiyon

WoS İndeksli Yayınlar Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu