WATER-TO-CEMENT RATIO PREDICTION USING ANNS FROM NON-DESTRUCTIVE AND CONTACTLESS MICROWAVE MEASUREMENTS

In concrete industry, there is a need for water-to-cement ratio (w/c) estimation of cement-based materials since the w/c ratio of cement mixtures is typically given at the batch plant, and this ratio, sometimes, is deliberately changed to have a more workable cement mixture. To meet the requirements of accurate w/c ratio determination of cement-based materials, in this research paper, we propose an artificial neural network approach for w/c ratio estimation of these materials using free-space non-contact reflection and transmission measurements of mortar specimens with w/c ratios of 0.40, 0.45, 0.50, 0.55 and 0.60. We have tested the network and observed less than 5 percent difference between the estimated and known values of w/c = 0.50.