Show simple item record

dc.contributor.authorBaghbani, Abolfazl
dc.contributor.authorShirani Faradonbeh, Roohollah
dc.contributor.authorAbuel-Naga, Hossam
dc.contributor.authorCosta, Susanga
dc.contributor.authorAlmasoudi, Rayed
dc.date.accessioned2023-04-19T03:09:46Z
dc.date.available2023-04-19T03:09:46Z
dc.date.issued2023
dc.identifier.citationBaghbani, A. and Shirani Faradonbeh, R. and Abuel-Naga, H. and Costa, S. and Almasoudi, R. 2023. Ultrasonic Characterization of Compacted Salty Kaolin–Sand Mixtures Under Nearly Zero Vertical Stress Using Experimental Study and Machine Learning. Geotechnical and Geological Engineering: an international journal.
dc.identifier.urihttp://hdl.handle.net/20.500.11937/91546
dc.identifier.doi10.1007/s10706-023-02441-5
dc.description.abstract

Understanding the dynamic behavior of soil is crucial for developing effective mitigation strategies for natural hazards such as earthquakes, landslides, and soil liquefaction, which can cause significant damage and loss of life. The ultrasonic wave testing method provides a non-invasive and reliable way of measuring the shear modulus, damping ratio and density of soils, which are fundamental parameters for understanding soil’s dynamic characteristics. The aim of this study was to investigate the effects of environmental factors, such as water salinity, soil liquid limit, plasticity index, dry density, and water content, on ultrasonic wave velocities (specifically shear and primary waves) in kaolin–sand mixtures subjected to near-zero vertical stress, as well as to predict these effects utilizing two unique artificial intelligence methods, including Classification and Regression Random Forests (CRRF) and Artificial Neural Networks (ANN), which, to our knowledge, have not been utilized in previous literature. The CRRF and ANN models were developed using two well-known algorithms and five different architectures using a database of 128 datasets. Water salinity, dry density, water content, liquid limit and plasticity index were predictor variables. The results showed that both CRRF and ANN were highly accurate. The coefficient of determination (R2) and mean absolute error (MAE) of the best CRRF were 0.963 and 9.191, respectively to predict Vs, and 0.974 and 7.809 to predict Vp, respectively. Furthermore, in ANN, R2 and MAE were respectively 0.994 and 0.016 to predict both Vs and Vp. According to importance analysis, liquid limit, molality, and dry density are the most critical parameters, while water content is the least critical.

dc.publisherSpringer Nature
dc.titleUltrasonic Characterization of Compacted Salty Kaolin–Sand Mixtures Under Nearly Zero Vertical Stress Using Experimental Study and Machine Learning
dc.typeJournal Article
dcterms.source.issn0960-3182
dcterms.source.titleGeotechnical and Geological Engineering: an international journal
dc.date.updated2023-04-19T03:09:22Z
curtin.departmentWASM: Minerals, Energy and Chemical Engineering
curtin.accessStatusFulltext not available
curtin.facultyFaculty of Science and Engineering
curtin.contributor.orcidShirani Faradonbeh, Roohollah [0000-0002-1518-3597]
curtin.contributor.scopusauthoridShirani Faradonbeh, Roohollah [56598081500]
curtin.repositoryagreementV3


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record