Computational & Technology Resources
an online resource for computational,
engineering & technology publications
ADVANCES IN ENGINEERING COMPUTATIONAL TECHNOLOGY
Edited by: B.H.V. Topping
Prediction of Strength for Concrete Specimens using Artificial Neural Networks
A. Kaveh and A. Khalegi
Iran University of Technology, Narmak, Tehran, Iran, Building and Housing Research Centre, Tehran, Iran
A. Kaveh, A. Khalegi, "Prediction of Strength for Concrete Specimens using Artificial Neural Networks", in B.H.V. Topping, (Editor), "Advances in Engineering Computational Technology", Civil-Comp Press, Edinburgh, UK, pp 165-171, 1998. doi:10.4203/ccp.53.4.3
Keywords: concrete strength, admixture, artificial neural networks, backpropagation algorithm.
Artificial Neural Networks are trained for different types of concrete mixtures, in order to predict the 7-day and 28-day strength of concrete specimens. Both plain and admixture concretes are considered. Employing the Backpropagation algorithm, neural nets with one, two and three hidden layers are trained and compared. The most efficient networks are then selected and used for predicting the strength of concrete mixtures, with reasonably small errors.
purchase the full-text of this paper (price £20)