In a previous research evaluating the effectiveness of fly ash to mitigate Alkali Silica Reaction (ASR), a novel approach of combined NIST model and ASTM C 311 test was developed to determine pore solution concentration (PSC) of concrete. In this research, a thermodynamic based GEMS model accounting for alkali binding was developed to predict pore solution of binary and ternary concrete mixes. Results demonstrate that crystalline alkali sulfate phases (such as Theradnite, Arcanite, etc.) in cement and fly ash are principal sources of available alkalis in the pore solution of concrete, which may or may not have a bearing with total bulk alkali phases Na2O and K2O. A distinct advantage of the combined NIST and ASTM C 311 approach over the traditional NIST model lies in its ability to account for soluble alkali fractions from fly ash (available alkalis) as opposed to bulk fractions. Results from thermodynamic modeling validate this novel approach for the reliable determination of PSC. To further simplify this proposed approach, the development of a machine learning model is in progress to predict available alkalis from fly ash based on bulk chemical composition. The objective is to present a simplified approach for industry practitioners and researchers to predict PSC of concrete, with acceptable accuracy, based on its bulk composition. The combined application of two approaches provides a reliable method to evaluate PSC and determine the optimum level of Fly Ash replacement to mitigate ASR, based on the aggregate’s threshold alkalinity (TA) - PSC relationship.
Details
| Title | Linking Pore Solution Chemistry of Concrete to ASR Potential Through Machine Learning |
| Duration | 20 Mins |
| Language | English |
| Format | MP4 |
| Size | 54 MB |
| Download Method | Direct Download |
| Download Links | BECOME A MEMBER VIEW DOWNLOAD LINKS |
The post Linking Pore Solution Chemistry of Concrete to ASR Potential Through Machine Learning appeared first on Civil Engineering Community.
source https://www.civilax.com/linking-pore-solution-chemistry-of-concrete-to-asr-potential-through-machine-learning/
