Predicting the compressive strength of self-compacting concrete using Elman artificial neural network with two different sets of input parameters

In recent years, artificial neural networks converted from a theoretical approach to the widely-used technology with successful applications to different problems. In fact, artificial neural networks are a powerful tool that give appropriate solutions to problems which are difficult to solve through...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Atefeh Gholamzadeh Chitgar, Javad Berenjian
Formato: article
Lenguaje:FA
Publicado: Iranian Society of Structrual Engineering (ISSE) 2019
Materias:
Acceso en línea:https://doaj.org/article/da163dab9cc5401aace7e208fe308f59
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Descripción
Sumario:In recent years, artificial neural networks converted from a theoretical approach to the widely-used technology with successful applications to different problems. In fact, artificial neural networks are a powerful tool that give appropriate solutions to problems which are difficult to solve through conventional techniques. Nowadays, these networks, which are inspired by the biological nervous system, are also extensively used to solve a wide range of complex problems in civil engineering. The purpose of the current study is a performance evaluation of the Elman artificial neural networks with various input parameters in order to predict the compressive strength of Self Compacting Concrete (SCC). Therefore, once, 8 effective parameters and next, in order to simulate a real experimental conditions, 140 parameters were entered as inputs in the Elman neural networks. According to the results, Elman neural networks, as a reliable tool, have high strength for predicting the desired properties along with saving time and cost. In addition, in both 7 and 28-day compressive strength, the constructed networks with 140 input parameters compared to ones with 8, have 74.54 and 70.44 percent improvement respectively regarding their test errors. The effective inputs straightly affect the networks errors in the prediction of the desired properties.