دورية
Advanced regression models for assessing the strength of multi-walled carbon nanotube-modified high-volume fly ash concrete
| العنوان: | Advanced regression models for assessing the strength of multi-walled carbon nanotube-modified high-volume fly ash concrete |
|---|---|
| المؤلفون: | Kumar, Anish, Sinha, Sanjeev, Pandey, Divya, Maurya, Madan Chandra, Chauhan, Vinay Bhushan |
| المصدر: | Asian Journal of Civil Engineering; 20230101, Issue: Preprints p1-22, 22p |
| مستخلص: | This study presents a comprehensive analysis of the strength, durability, and microstructural characteristics of high-volume fly ash (HVFA) concrete incorporating varying levels of multi-walled carbon nanotubes (MWCNTs), ranging from 0% to 2% by weight. Machine learning-based regression techniques, specifically LASSO, RIDGE, and ENET, were employed to construct predictive models for split tensile strength and flexural strength of HVFA concrete. The investigation determined that the optimal fly ash content for HVFA concrete was 55%. To ensure effective dispersion of MWCNTs within the aqueous medium, ultrasonication was applied. Remarkably, normal concrete (NC) specimens doped with 2% MWCNT exhibited the highest compressive, split tensile, and flexural strength, while HVFA samples displayed comparatively lower strength properties. In the face of a 90-day exposure to a corrosive environment (4% H2SO4), NC samples containing 2% MWCNT demonstrated the least disintegration. All specimens examined in this study achieved classifications of "good" or "excellent" based on the results obtained from ultrasonic pulse velocity tests. The inclusion of MWCNTs did not significantly impact the density of the concrete samples. Furthermore, the water permeability of HVFA concrete decreased, and the highest values in the Rapid Chloride Penetration Test (RCPT) were observed for HVFA samples. The introduction of MWCNTs into the matrix led to a decline in RCPT values. Microscopic analysis of the samples confirmed the presence of MWCNT clusters and the formation of calcium silicate hydrate (CSH) gel. Among the regression models developed, the ENET model for split tensile strength exhibited the highest performance with an R2value of 0.962, surpassing LASSO and RIDGE regression models. Notably, the R2value of LASSO and ENET was identical, measuring 0.871. These regression modelling outcomes were consistently supported by correlation plots. |
| قاعدة البيانات: | Supplemental Index |
Find this article in full text from Springer Verlag. | تدمد: | 15630854 2522011X |
|---|---|
| DOI: | 10.1007/s42107-023-00906-9 |