Neural modelling of the depth of wear determined using the rotating-cutter method for concrete with a high volume of high-calcium fly ash

The application of concrete with a high volume (over 30% of the cement mass) of high-calcium fly ash has recently gained more attention. This solution is more and more frequently used, especially in the case of horizontally formed elements (e.g. slabs, overlays), and it may contribute to their lower abrasion resistance. Traditional abrasion resistance evaluations, although successful, are based on destructive test methods. Samples not only have to be damaged, but also tested in laboratory conditions. The article presents an artificial neural network that was constructed, based on statistical and numerical analysis, for the purpose of developing a precise, non-destructive abrasion resistance testing method. The choice of the appropriate inputs, and also the best algorithm with the right number of neurons, resulted in very high accuracy and minor errors during the prediction process. The greatest undervaluing of the predicted results in relation to the conducted measurements was equal to 0.35 mm, while the greatest overstating was equal to 0.22 mm. The authors conclude that it is possible to predict the depth of wear based on the components and age of concrete, as well as the time of testing.