Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms
Concrete is the most utilized material (i.e., average production of 2 billion tons per year) for the construction of buildings, bridges, roads, dams, and several other important infrastructures. The strength and durability of these structures largely depend on the compressive strength of the concret...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
Elsevier Ltd
2024
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1833413767330791424 |
|---|---|
| author | Ziyad Sami B.H. Ziyad Sami B.F. Kumar P. Ahmed A.N. Amieghemen G.E. Sherif M.M. El-Shafie A. |
| author2 | 57481263600 |
| author_facet | 57481263600 Ziyad Sami B.H. Ziyad Sami B.F. Kumar P. Ahmed A.N. Amieghemen G.E. Sherif M.M. El-Shafie A. |
| author_sort | Ziyad Sami B.H. |
| building | UNITEN Library |
| collection | Institutional Repository |
| content_provider | Universiti Tenaga Nasional |
| content_source | UNITEN Institutional Repository |
| continent | Asia |
| country | Malaysia |
| description | Concrete is the most utilized material (i.e., average production of 2 billion tons per year) for the construction of buildings, bridges, roads, dams, and several other important infrastructures. The strength and durability of these structures largely depend on the compressive strength of the concrete. The compressive strength of concrete depends on the proportionality of the key constituents (i.e., fine aggregate, coarse aggregate, cement, and water). However, the optimization of the constituent proportions (i.e., matrix design) to achieve high-strength concrete is a challenging task. Furthermore, it is essential to reduce the carbon footprint of the cementitious composites through the optimization of the matrix. In this research, machine learning algorithms including regression models, tree regression models, support vector regression (SVR), ensemble regression (ER), and gaussian process regression (GPR) were utilized to predict the compressive and tensile concrete strength. Also, the model performance was characterized based on the number of input variables utilized. The dataset used in this research was compiled from journal publications. The results showed that the exponential GPR had the highest performance and accuracy. The model had an impressive performance during the training phase, with a R2 of 0.98, RMSE of 2.412 MPa, and MAE of 1.6249 MPa when using 8 input variables to predict the compressive strength of concrete. In the testing phase, the model maintained its accuracy with a R2 of 0.99, RMSE of 0.0025134 MPa, and MAE of 0.0016367 MPa. In the training and testing phases, the exponential GPR also demonstrated high accuracy in predicting the tensile strength with an R2, RMSE, and MAE of 0.99, 0.00049247 MPa, and 0.00036929 MPa, respectively. Furthermore, in the prediction of tensile strength the number of variables utilized had an insignificant effect on the performance of the models. However, in predicting the compressive strength, an increase in the number of input variables lead to an enhancement in the performance metrics. The results of this research can allow for the quick and accurate prediction of the strength of a given concrete mixture design. � 2023 |
| format | Article |
| id | my.uniten.dspace-34183 |
| institution | Universiti Tenaga Nasional |
| publishDate | 2024 |
| publisher | Elsevier Ltd |
| record_format | dspace |
| spelling | my.uniten.dspace-341832024-10-14T11:18:19Z Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms Ziyad Sami B.H. Ziyad Sami B.F. Kumar P. Ahmed A.N. Amieghemen G.E. Sherif M.M. El-Shafie A. 57481263600 57481286200 57206939156 57214837520 58089424300 57188776821 16068189400 Compressive strength Concrete Machine learning Predictive modeling Tensile strength Bridges Carbon footprint Concrete aggregates Concrete mixtures Forecasting High performance concrete Learning algorithms Machine learning Matrix algebra Regression analysis Structural design Tensile strength Tensile testing Compressive strength of concrete Concrete Gaussian process regression Input variables Machine learning algorithms Machine-learning Optimisations Performance Predictive models Regression modelling Compressive strength Concrete is the most utilized material (i.e., average production of 2 billion tons per year) for the construction of buildings, bridges, roads, dams, and several other important infrastructures. The strength and durability of these structures largely depend on the compressive strength of the concrete. The compressive strength of concrete depends on the proportionality of the key constituents (i.e., fine aggregate, coarse aggregate, cement, and water). However, the optimization of the constituent proportions (i.e., matrix design) to achieve high-strength concrete is a challenging task. Furthermore, it is essential to reduce the carbon footprint of the cementitious composites through the optimization of the matrix. In this research, machine learning algorithms including regression models, tree regression models, support vector regression (SVR), ensemble regression (ER), and gaussian process regression (GPR) were utilized to predict the compressive and tensile concrete strength. Also, the model performance was characterized based on the number of input variables utilized. The dataset used in this research was compiled from journal publications. The results showed that the exponential GPR had the highest performance and accuracy. The model had an impressive performance during the training phase, with a R2 of 0.98, RMSE of 2.412 MPa, and MAE of 1.6249 MPa when using 8 input variables to predict the compressive strength of concrete. In the testing phase, the model maintained its accuracy with a R2 of 0.99, RMSE of 0.0025134 MPa, and MAE of 0.0016367 MPa. In the training and testing phases, the exponential GPR also demonstrated high accuracy in predicting the tensile strength with an R2, RMSE, and MAE of 0.99, 0.00049247 MPa, and 0.00036929 MPa, respectively. Furthermore, in the prediction of tensile strength the number of variables utilized had an insignificant effect on the performance of the models. However, in predicting the compressive strength, an increase in the number of input variables lead to an enhancement in the performance metrics. The results of this research can allow for the quick and accurate prediction of the strength of a given concrete mixture design. � 2023 Final 2024-10-14T03:18:19Z 2024-10-14T03:18:19Z 2023 Article 10.1016/j.cscm.2023.e01893 2-s2.0-85147332991 https://www.scopus.com/inward/record.uri?eid=2-s2.0-85147332991&doi=10.1016%2fj.cscm.2023.e01893&partnerID=40&md5=d5407469ce8ad7b2762dc5b632fc69ff https://irepository.uniten.edu.my/handle/123456789/34183 18 e01893 All Open Access Gold Open Access Elsevier Ltd Scopus |
| spellingShingle | Compressive strength Concrete Machine learning Predictive modeling Tensile strength Bridges Carbon footprint Concrete aggregates Concrete mixtures Forecasting High performance concrete Learning algorithms Machine learning Matrix algebra Regression analysis Structural design Tensile strength Tensile testing Compressive strength of concrete Concrete Gaussian process regression Input variables Machine learning algorithms Machine-learning Optimisations Performance Predictive models Regression modelling Compressive strength Ziyad Sami B.H. Ziyad Sami B.F. Kumar P. Ahmed A.N. Amieghemen G.E. Sherif M.M. El-Shafie A. Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| title | Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| title_full | Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| title_fullStr | Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| title_full_unstemmed | Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| title_short | Feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| title_sort | feasibility analysis for predicting the compressive and tensile strength of concrete using machine learning algorithms |
| topic | Compressive strength Concrete Machine learning Predictive modeling Tensile strength Bridges Carbon footprint Concrete aggregates Concrete mixtures Forecasting High performance concrete Learning algorithms Machine learning Matrix algebra Regression analysis Structural design Tensile strength Tensile testing Compressive strength of concrete Concrete Gaussian process regression Input variables Machine learning algorithms Machine-learning Optimisations Performance Predictive models Regression modelling Compressive strength |
| url_provider | http://dspace.uniten.edu.my/ |