Simulation and prediction for the effect of natural and steel fibers on the performance of concrete using experimental analyses and artificial neural networks numerical modeling

Abstract

Utilization of fibers as concrete reinforcement is one method of preserving environment through the recycling of agriculture and industrial wastes. However, the optimum and efficient application of these fibers requires a measurable evaluation of their effect on properties and performance of concrete. In this study, the application of metallic steel fibers and natural (Linen) fibers in concrete industry is investigated, experimentally. Twenty one mixes are made with different mix proportions and with different types of fibers. The parameters are concrete characteristic strength, age and type of reinforcing fibers. Compressive, tensile and impact strength were measured for all mixes in addition to the residual compressive strength after exposure to elevated temperature. Measurements were also made using two Non Destructive Testing (NDT) techniques. Experimental results were compared and showed the enhancement level obtained by including steel and natural fibers. Following this experimental effort, one of the artificial intelligence techniques (Artificial Neural Network) was applied for simulating and predicting the performance of concrete with different mix proportions. The current paper introduced the Artificial Neural Network (ANN) technique to investigate the effect of natural and steel fibers on the performance of concrete. The results of this study showed that the ANN method with less effort was very efficiently capable of simulating and predicting the performance of concrete with different mix proportions and different types of fibers.

Modelagem preditiva de propriedades mecânicas em concretos reforçados com fibra de aço utilizando redes neurais artificiais

“…Classicamente, com o intuito de mitigar essas e outras limitações, o concreto é reforçado com barras de aço e, por essa razão, recebe a denominação de concreto armado (CA). Dessa forma, o aprimoramento de suas propriedades tornase necessário para proporcionar um material mais eficiente (HODHOD;ABDEEN, 2011).…”

“…Em decorrência das dificuldades mencionadas, a inteligência artificial surge como uma ferramenta que pode facilitar a redução do esforço na etapa de dosagem, economizando tempo e dinheiro, ao mesmo tempo em que se preserva a precisão dos resultados (HODHOD;ABDEEN, 2011). Entre as técnicas de Inteligência Artificial, existem as de aprendizado de máquina que podem fornecer uma solução para o problema de previsão de resistências do CRFA (KANG; YOO; GUPTA, 2021).…”