Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Unconfined Compressive Strength Prediction of Soils Improved with Biopolymers: Machine Learning Approach
https://orcid.org/http://orcid.org/0000-0002-7935-1334
Biopolymers act as alternative materials for traditional soil stabilizers. There is a numerical relationship to estimate the strength behavior of stabilized soil. White-box methods such as response surface (RSM) and gene expression programming (GEP) were used to estimate the unconfined compressive strength (UCS) of soil stabilized with xanthan gum. For this purpose, 248 laboratory data were collected and used. The effect of input parameters, such as plasticity index, xanthan gum content, curing time, curing temperature, initial water moisture content, and UCS of untreated soil, was investigated. The proposed RSM and GEP relationship’s R2, mean squared error (MSE), and mean absolute error (MAE) values were equal to 0.939, 0.6576, 0.0675, 0.3855, and 0.1926, 0.449, respectively. According to the statistical coefficient, the results of RSM are more accurate than those of GEP. The sensitivity analysis showed that initial water moisture content and UCS of untreated soil parameters were critical for both methods. Also, curing temperature in the DOE method significantly affects the UCS changes, while this parameter is not essential in the GEP method. The comparison of the proposed machine learning models with laboratory results demonstrated the satisfactory performance of RSM and GEP in estimating the UCS of stabilized soil samples.
Unconfined Compressive Strength Prediction of Soils Improved with Biopolymers: Machine Learning Approach
https://orcid.org/http://orcid.org/0000-0002-7935-1334
Biopolymers act as alternative materials for traditional soil stabilizers. There is a numerical relationship to estimate the strength behavior of stabilized soil. White-box methods such as response surface (RSM) and gene expression programming (GEP) were used to estimate the unconfined compressive strength (UCS) of soil stabilized with xanthan gum. For this purpose, 248 laboratory data were collected and used. The effect of input parameters, such as plasticity index, xanthan gum content, curing time, curing temperature, initial water moisture content, and UCS of untreated soil, was investigated. The proposed RSM and GEP relationship’s R2, mean squared error (MSE), and mean absolute error (MAE) values were equal to 0.939, 0.6576, 0.0675, 0.3855, and 0.1926, 0.449, respectively. According to the statistical coefficient, the results of RSM are more accurate than those of GEP. The sensitivity analysis showed that initial water moisture content and UCS of untreated soil parameters were critical for both methods. Also, curing temperature in the DOE method significantly affects the UCS changes, while this parameter is not essential in the GEP method. The comparison of the proposed machine learning models with laboratory results demonstrated the satisfactory performance of RSM and GEP in estimating the UCS of stabilized soil samples.
Unconfined Compressive Strength Prediction of Soils Improved with Biopolymers: Machine Learning Approach
https://orcid.org/http://orcid.org/0000-0002-7935-1334
Biopolymers act as alternative materials for traditional soil stabilizers. There is a numerical relationship to estimate the strength behavior of stabilized soil. White-box methods such as response surface (RSM) and gene expression programming (GEP) were used to estimate the unconfined compressive strength (UCS) of soil stabilized with xanthan gum. For this purpose, 248 laboratory data were collected and used. The effect of input parameters, such as plasticity index, xanthan gum content, curing time, curing temperature, initial water moisture content, and UCS of untreated soil, was investigated. The proposed RSM and GEP relationship’s R2, mean squared error (MSE), and mean absolute error (MAE) values were equal to 0.939, 0.6576, 0.0675, 0.3855, and 0.1926, 0.449, respectively. According to the statistical coefficient, the results of RSM are more accurate than those of GEP. The sensitivity analysis showed that initial water moisture content and UCS of untreated soil parameters were critical for both methods. Also, curing temperature in the DOE method significantly affects the UCS changes, while this parameter is not essential in the GEP method. The comparison of the proposed machine learning models with laboratory results demonstrated the satisfactory performance of RSM and GEP in estimating the UCS of stabilized soil samples.
Unconfined Compressive Strength Prediction of Soils Improved with Biopolymers: Machine Learning Approach
Transp. Infrastruct. Geotech.
Ghazavi, Mahmoud (Autor:in) / Afrakoti, Mobina Taslimi Paein (Autor:in)
01.01.2025
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Prediction of unconfined compressive strength of lime treated soils
| Taylor & Francis Verlag | 2024
| British Library Online Contents | 2017
Prediction of Unconfined Compressive Strength of Stabilized Sand Using Machine Learning Methods
| Springer Verlag | 2025