Application of an artificial neural network for predicting compressive and flexural strength of basalt fiber added lightweight concrete

• Published in: Challenge Journal of Concrete Research Letters Vol. 12; no. 1; p. 12
• Main Authors: Calis, Gokhan, Yıldızel, Sadık Alper, Keskin, Ülkü Sultan
• Format: Journal Article
• Language: English
• Published: 12.03.2021
• ISSN: 2548-0928; 2548-0928
• DOI: 10.20528/cjcrl.2021.01.002

Concrete is known as one of the fundamental materials in construction with its high amount of use. Lightweight concrete (LWC) can be a good alternative in reducing the environmental effect of concrete by decreasing the self-weight and dimensions of the structure. In order to reduce self-weight of concrete artificial aggregates, some of which are produced from waste materials, are utilized, and it also contributes to develop a sustainable material Artificial neural networks have been the focus of many scholars for long time with the purpose of analyzing and predicting the lightweight concrete compressive and flexural strengths. The artificial neural network is more powerful method in terms of providing explanation and prediction in engineering studies. It is proved that the error rate of ANN is smaller than regression method. Furthermore, ANN has superior performance over nonlinear regression model. In this paper, an ANN based system is proposed in order to provide a better understanding of basalt fiber reinforced lightweight concrete. In the regression analysis predicted vs. experimental flexural strength, R-sqr is determined to be 86%. The most important strength contributing factors were analyzed within the scope of this study.