Keywords: artificial neural network, blended systems, high strength concrete, mercury intrusion porosimetry, microsilica, pore structure, pulverised fuel ash.

The pore structure of paste, mortar or concrete affects the behaviour of the material and can give insight into both material microstructure and its performance. Permeability and other durability related properties of concrete are directly affected by the pore structure. Therefore, the size of the pores, their extent and shape are factors of paramount importance in deciding the suitability of materials for practical applications. Conventional concrete often fails to prevent the intrusion of moisture and aggressive ions adequately, therefore, special concrete with a low permeability is needed. The use of supplementary cementing materials such as pulverised fuel ash (PFA), microsilica have been reported to refine the pore structure which results reduction in permeability and reduces aggression of chemicals such as chlorides [1,2]. In this study, the pore structure of cement mortar incorporating PFA ash up to 40% and microsilica up to 15% as partial cement replacements was conducted using mercury intrusion porosimetry [3]. It was observed that PFA incorporation has slight pore refinement effect whilst microsilica inclusion demonstrated significant improvement on pore refinement.

In this investigation, an artificial neural network (ANN) based on the radial basis function (RBF) have been used. A RBF neural network is a layered network consisting of an input layer, an output layer and at least one layer of nonlinear processing elements known as the hidden layer [4,5]. The network developed in this investigation has eight units in the input layer and two units in the output layer. The experimentally obtained data have been divided into two sets, one for the network learning called learning set, and the other for testing the network called testing set.

Based on the experimentally obtained results, an ANN has been used to establish its applicability for the prediction of the pore structure and pore size, small pores (<0.1 µm) and large pores (>0.1 µm), of concrete. It has been observed that there is a good correlation between experimental values and those predicted using an ANN. Therefore, it is possible to predict the pore structure and pore size of concrete having using ANN.

1

V.M. Malhotra and P.K. Mehta, "Pozzolanic and cementitious materials - Advances in concrete technology", Vol. I, Gordon and Breach, Netherlands, 1996.

2

FIP Report, "Condensed silica fume in concrete. FIP state-of-art report", FIP commission of Concrete, Thomas Telford House, London, pp. 37, 1988.

3

Micromeritics Manual, Micromeritics manual instruction for mercury porosimetry model Auto Pore Sizer 9320, 1982.

4

I.J. Leontaritis and S.A. Billings, "Input-output parametric models for nonlinear systems part-1: deterministic nonlinear systems", International Journal of Control, 41, 303-328, 1985. doi:10.1080/0020718508961129

5

S. Chen, C.F.N. Cowan and P.M. Grant, "Orthogonal Least Squares Learning Algorithm for Radial Basis Function Networks", IEEE Trans. on Neural Networks, 2(2), 302-309, 1991. doi:10.1109/72.80341

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £62 +P&P)