Application of soft computing methods for predicting the elastic modulus of recycled aggregate concrete

摘要

The use of recycled concrete aggregate (RCA) as a replacement for natural aggregate in concrete mixtures provides a wide variety of benefits such as reduced cost and pollution, reduced CO2 footprint, and reduced pressure on natural resources. Due to the inferior properties of the RCA compared to the natural one, the mechanical properties of the concretes made with former are different than those of the concretes made with latter. Therefore, understanding the relationship between the mechanical properties and the mix proportions of recycled aggregate concrete (RAC) is essential before widely adopting this material by construction industry. Elastic modulus of concrete is one of the major mechanical properties used by civil engineers in design applications to calculate the drift and deflection of the concrete structures. In this study, four types of soft computing methods, namely, artificial neural network (ANN), fuzzy TSK, support vector regression (SVR) and radial basis function neural network (RBFNN) were employed to predict the 28-day elastic modulus of RAC (E-RAc). To develop the predictive models of ERK using these soft computing techniques, a comprehensive dataset containing 400 RAC mix design records was collected from internationally published literature. The results indicated that all the proposed models are successfully able to predict the E-RAc close agreement with the experimental results. Moreover, error measures were used to compare the performance of the soft computing techniques with each other. The results showed that the proposed models based on the SVR and ANN techniques outperform the models proposed using other techniques. (C) 2017 Elsevier Ltd. All rights reserved.