Master'sOpen Access

Determination of mix designs of lightweight concrete produced with different materials using machine learning methods

2019
0 views
0 downloads
Advisor: Doç. Dr. Kürşat Esat Alyamaç

Abstract (EN)

The use of lightweight concrete is consistently increasing nowadays. This kind of concrete, which has many advantages, is classified based on its strength. Additionally, lightweight concretes are generally produced as structural lightweight concrete, semi-structural concrete and insulation concretes. It has been seen that different types of lightweight aggregates are used in different lightweight concrete applications. While lightweight aggregates such as fly ash and expanded clay are mostly used in lightweight concrete production, pumice stone and volcanic slag aggregates are used in the production of semi-lightweight concrete. Perlite and vermiculite are aggregates with high heat and sound insulation. Therefore, perlite and vermiculite aggregates are used in insulation concrete production. In this study, a theoretical research has been performed on the mix design of lightweight concrete. In this study, the current mixture design of lightweight concretes produced by using different lightweight aggregates in the literature were examined. The amounts of the mix components, the compressive strength, dry density values were taken from the appropriate studies. Afterwards, mix design was developed using statistical analysis and response surface method from the recorded data. The four different types of artificial neural networks are designed. The targets in ANN 1, ANN 2, ANN 3 types were determined as dry density, slump and compressive strength, respectively. In ANN 4 type, it is aimed to estimate the amount of lightweight aggregate by adding target strength instead of lightweight aggregate. These mix designs were compared with the results obtained in the literature. The correlation value obtained with the artificial neural network was higher than the response surface method. The differences between the results obtained with ANN and RSM and the data collected were found within the appropriate limits. As a result, these mix designs will provide a practical mix design for lightweight concretes to be produced with diffirent lightweight aggregate. Therefore, it will contribute to further study of lightweight concrete in the literature.

Author

Rabia Nur Aydın Sağlam

How to Cite

Rabia Nur Aydın Sağlam (Master Thesis). Determination of mix designs of lightweight concrete produced with different materials using machine learning methods, 2019, Fırat University.

Keywords

License

Tüm Hakları Saklıdır

This work is shared under the specified license terms.

More theses from Fırat University