Study on the spatiotemporal evolution and prediction of internal porosity in concrete specimens under sulfate attack based on machine learning models
Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review
Author(s)
Related Research Unit(s)
Detail(s)
Original language | English |
---|---|
Article number | 103258 |
Journal / Publication | Results in Engineering |
Volume | 24 |
Online published | 31 Oct 2024 |
Publication status | Published - Dec 2024 |
Link(s)
DOI | DOI |
---|---|
Attachment(s) | Documents
Publisher's Copyright Statement
|
Link to Scopus | https://www.scopus.com/record/display.uri?eid=2-s2.0-85208936052&origin=recordpage |
Permanent Link | https://scholars.cityu.edu.hk/en/publications/publication(36bd4f35-1782-49f5-91e5-d045e28f1593).html |
Abstract
Concrete structures deteriorate due to sulfate attack, leading to changes in porosity. Understanding porosity evolution under sulfate attack helps improve concrete durability through protective coatings like silane. Existing research suggests that while silane can protect freshly placed concrete with good integrity, it may not be effective on defective freshly placed concrete structures and may even accelerate damage to already cracked concrete. The effect of silane on sulphate attacked but uncracked concrete structures is unknown. This study addressed this gap by conducting accelerated aging tests on concrete specimens from tunnel sites, subjected to sulfate solutions. CT scans analyzed porosity changes in silane-coated and uncoated control groups after 60, 120, and 180 days. The study revealed temporal porosity evolution through variation curves along the axial and radial directions of the specimens. Machine learning methods were employed to predict porosity changes based on CT data. The results demonstrated that CT scan analysis and image processing techniques can effectively visualize internal porosity and its spatiotemporal evolution. Silane coatings showed a significant protective effect on sulfate-exposed concrete. Random Forest Regression emerged as the most accurate model for predicting porosity changes, highlighting the potential of machine learning to address CT scan resolution limitations and reduce experimental costs. The findings of this paper provide important guidance and engineering application value for the safety protection of concrete structures in sulfate environments, especially for analyzing and preventing the deterioration of various cylindrical concrete structures (such as concrete piles) due to sulfate corrosion.
© 2024 The Authors. Published by Elsevier B.V.
© 2024 The Authors. Published by Elsevier B.V.
Research Area(s)
- Concrete corrosion, CT scan analysis, Machine learning models, Pore structures, Silane coatings, Sulphate attack
Citation Format(s)
Study on the spatiotemporal evolution and prediction of internal porosity in concrete specimens under sulfate attack based on machine learning models. / Gong, Chun; Liu, Dunwen; Cao, Kunpeng et al.
In: Results in Engineering, Vol. 24, 103258, 12.2024.
In: Results in Engineering, Vol. 24, 103258, 12.2024.
Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review
Download Statistics
No data available