Keywords: compressive strength, concrete, fine aggregate, granite powder, support vector machines.

This paper discusses the feasibility of using granite fines as a substitute for natural river-bed sand in concrete. A material that complements sand, thereby reducing the construction cost will be significant. Many researchers in the construction industries have identified some alternatives; namely quarry dust studied by Ilangovan and Nagamani in 2007 [1], rock dust studied by Nagaraj and Banu in 1996 [2], and crushed rock studied by Murdock et al. in 1991 [3]. Mix design was developed for M20 grade concrete using the IS:10262(2009) code and ACI design. Tests were conducted using 10 cm3 concrete cubes to measure the strength of concrete with the replacement of sand by granite fines. The percentage of granite powder added by weight was 0-100 % with 3.5 % intervals as a replacement for the sand used in concrete. The water-to-binder ratio used was 0.50. Results of the compressive test show that the inclusion of granite fines in the concrete mix has a significant influence on its compressive strength. Observation validates the 35% granite fines content as optimum for augmented structural performance. Compressive strengths evaluated for the seven day and twenty eight days were 34.4 and 46.8 N/mm² respectively.

This paper also incorporates support vector machines (SVM), a method for classification and regression derived from statistical learning theory by Vapnik [4], in predicting the compressive strength of concrete. The SVM originated from the concept of statistical learning theory pioneered by Boser et al. [5]. Our data set has thirty inputs through which 70% were randomly selected as a training data set and the remaining 30% as a testing data set. Each of the data set had been normalised to reduce the initial error and size of the data. The percentage replacement of sand with granite fines was considered as input, and the given output of compressive strengths were calculated using SVM. This study provides equations for the prediction of the compressive strength of the granite mixed concrete for the compressive strength of granite mixed concrete, thereby giving the performance of training and testing dataset for the compressive strength using SVM. The output predicted shares a correlation coefficient close to 1, thus proving in both the cases, i.e. seven-day and twenty eight-day compressive strengths, that the SVM developed has the ability to predict the compressive strength The results show the use of the support vector machine based model to effectively predict the compressive strength of high performance concrete.

1

R. Ilangovan, K. Nagamani, "Application of Quarry Dust in Concrete construction. High Performance Concrete", Federal Highway Administration, 1-3, 2007.

2

T.S. Nagaraj, Banu, Zahida, "Efficient utilization of rock dust and pebbles as aggregates in Portland cement concrete", Indian Concrete Journal, 70(I), 1, 4, 1996.

3

L.J. Murdock, K.M. Brook, J.D. Dewar, "Concrete Materials and Practice", Edward Arnold London, 1991.

4

B.E. Boser, I.M. Guyon, V.N. Vapnik, "A training algorithm for optimal margin classifiers", Proceedings of the Annual Conference on Computational Learning Theory, ACM Press, Pittsburgh, PA, 144-152, 1992. doi:10.1145/130385.130401

5

V.N. Vapnik, "Statistical Learning Theory", John Wiley and Sons, New York, 1998.

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 £50 +P&P)