Self-Organizing Maps for Durability Prediction of Fly Ash Geopolymer in Chloride Environment
| dc.contributor.author | Fong Wen Lee | |
| dc.date.accessioned | 2026-05-05T08:18:13Z | |
| dc.date.issued | 2026-05 | |
| dc.description | Geopolymer concrete offers a promising alternative to traditional Portland cement concrete, exhibiting comparable mechanical and durability performance, while reducing environmental impacts. However, its mechanical properties and durability depend on many factors such as the water/binder ratio, activator concentration, and curing temperature. This study utilized a Self-organizing Map (SOM), an unsupervised artificial neural network, to model experimental data to predict the factors controlling geopolymer concrete's durability in chloride environments. The research specifically investigated how variations in the water-to-binder ratio and activator molarity impact performance. The model was trained on data obtained from cylindrical samples that were heat-cured, matured for 28 days, and then tested for chloride migration. This modeling approach effectively highlighted patterns and correlations within the dataset, offering essential insights into the chloride environment. A key quantitative finding was the validation of the SOM and GSOM model's reliability, as cross-validation confirmed their high prediction accuracy and their clustering patterns closely corresponded with empirical trends. Based on the analysis of 2,268 data points from 9 different recipes, the study confirm that a mix of 12 M molarity and a 0.5 water-binder ratio offers the highest chloride resistance, providing a optimal recipe for durable geopolymer concrete. | |
| dc.identifier.uri | https://scholarhub.unimas.my/handle/123456789/639 | |
| dc.language.iso | English | |
| dc.publisher | Universiti Malaysia Sarawak | |
| dc.relation.ispartofseries | Faculty of Engineering | |
| dc.subject | Fly ash, Chlorides | |
| dc.title | Self-Organizing Maps for Durability Prediction of Fly Ash Geopolymer in Chloride Environment | |
| dc.type | Masters |