Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete
The environmental challenges associated with cement production have driven the development of alkali-activated concrete (AAC) as a sustainable alternative to Portland cement-based materials. AAC incorporates industrial byproducts, such as fly ash (FA) and ground granulated blast furnace slag (GGBFS)...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | Cleaner Engineering and Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666790825001442 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850118749917544448 |
|---|---|
| author | Dina A. Emarah |
| author_facet | Dina A. Emarah |
| author_sort | Dina A. Emarah |
| collection | DOAJ |
| description | The environmental challenges associated with cement production have driven the development of alkali-activated concrete (AAC) as a sustainable alternative to Portland cement-based materials. AAC incorporates industrial byproducts, such as fly ash (FA) and ground granulated blast furnace slag (GGBFS), as binders, significantly reducing carbon emissions. While many studies have explored predictive modeling of AAC's compressive strength (CS), this study stands out by addressing critical gaps in the field. Using a comprehensive dataset of 1590 samples with 14 input variables, it captures the complex, multi-variable dependencies affecting AAC's mechanical behavior. Unlike previous studies that often focus on a limited set of parameters or single-variable models, this work evaluates and compares four advanced predictive models: Linear Regression (LR), Multi-Linear Regression (MLR), Non-Linear Regression (NLR), and Artificial Neural Networks (ANN). The ANN model, with its ability to handle non-linear interactions, significantly outperformed traditional methods, achieving the highest Coefficient of Determination (R2 = 0.96) and the lowest Root Mean Squared Error (RMSE = 2.82 MPa). Moreover, this study introduces a sensitivity comparison that was carried out for the ANN model to discover and analyze the most critical input parameter that influences the CS of AAC. The results reveal that curing temperature is the most influential factor, followed by the alkaline solution-to-binder (AL/b) ratio, sodium hydroxide concentration, and specimen age. Additionally, the study introduced advanced performance metrics, including R2, RMSE, Scatter Index (SI), Objective Function (OBJ), and Scatter Index (SI), to provide a more robust validation of the models. By incorporating diverse parameters, employing advanced machine learning techniques, and performing a comprehensive sensitivity analysis, this research establishes a new benchmark for predictive modeling of AAC's CS. The findings offer actionable insights for optimizing AAC formulations and further support the broader adoption of AAC as an eco-friendly construction material. In addition, this study lays the foundation for future innovations, including hybrid modeling approaches and sustainability-focused assessments. |
| format | Article |
| id | doaj-art-99fe061490e8438dac4424a482b7e41f |
| institution | OA Journals |
| issn | 2666-7908 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Cleaner Engineering and Technology |
| spelling | doaj-art-99fe061490e8438dac4424a482b7e41f2025-08-20T02:35:47ZengElsevierCleaner Engineering and Technology2666-79082025-07-012710102110.1016/j.clet.2025.101021Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concreteDina A. Emarah0Construction Research Institute (CRI), National Water Research Center (NWRC), Delta-Barrage, 13621, EgyptThe environmental challenges associated with cement production have driven the development of alkali-activated concrete (AAC) as a sustainable alternative to Portland cement-based materials. AAC incorporates industrial byproducts, such as fly ash (FA) and ground granulated blast furnace slag (GGBFS), as binders, significantly reducing carbon emissions. While many studies have explored predictive modeling of AAC's compressive strength (CS), this study stands out by addressing critical gaps in the field. Using a comprehensive dataset of 1590 samples with 14 input variables, it captures the complex, multi-variable dependencies affecting AAC's mechanical behavior. Unlike previous studies that often focus on a limited set of parameters or single-variable models, this work evaluates and compares four advanced predictive models: Linear Regression (LR), Multi-Linear Regression (MLR), Non-Linear Regression (NLR), and Artificial Neural Networks (ANN). The ANN model, with its ability to handle non-linear interactions, significantly outperformed traditional methods, achieving the highest Coefficient of Determination (R2 = 0.96) and the lowest Root Mean Squared Error (RMSE = 2.82 MPa). Moreover, this study introduces a sensitivity comparison that was carried out for the ANN model to discover and analyze the most critical input parameter that influences the CS of AAC. The results reveal that curing temperature is the most influential factor, followed by the alkaline solution-to-binder (AL/b) ratio, sodium hydroxide concentration, and specimen age. Additionally, the study introduced advanced performance metrics, including R2, RMSE, Scatter Index (SI), Objective Function (OBJ), and Scatter Index (SI), to provide a more robust validation of the models. By incorporating diverse parameters, employing advanced machine learning techniques, and performing a comprehensive sensitivity analysis, this research establishes a new benchmark for predictive modeling of AAC's CS. The findings offer actionable insights for optimizing AAC formulations and further support the broader adoption of AAC as an eco-friendly construction material. In addition, this study lays the foundation for future innovations, including hybrid modeling approaches and sustainability-focused assessments.http://www.sciencedirect.com/science/article/pii/S2666790825001442Fly ashGGBFSAlkali-activated concreteCompressive strengthMix proportion modeling |
| spellingShingle | Dina A. Emarah Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete Cleaner Engineering and Technology Fly ash GGBFS Alkali-activated concrete Compressive strength Mix proportion modeling |
| title | Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete |
| title_full | Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete |
| title_fullStr | Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete |
| title_full_unstemmed | Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete |
| title_short | Multivariate predictive modeling of compressive strength in ground granulated blast furnace slag/fly ash-based alkali-activated concrete |
| title_sort | multivariate predictive modeling of compressive strength in ground granulated blast furnace slag fly ash based alkali activated concrete |
| topic | Fly ash GGBFS Alkali-activated concrete Compressive strength Mix proportion modeling |
| url | http://www.sciencedirect.com/science/article/pii/S2666790825001442 |
| work_keys_str_mv | AT dinaaemarah multivariatepredictivemodelingofcompressivestrengthingroundgranulatedblastfurnaceslagflyashbasedalkaliactivatedconcrete |