ANN prediction of cement mortar compressive strength, influence of cement strength class

Abstract

An artificial neural network (ANN) study is presented to predict the compressive strength (Fc) of mortar mixtures containing different cement strength classes of CME 32.5, 42.5, and 52.5MPa. For this purpose, 54 mixtures considering six water/cement ratios (W/C) (0.25, 0.3, 0.35, 0.4, 0.45, and 0.5) and three sand/cement ratios (S/C) (2.5, 2.75, and 3) along with the abovementioned three types of cement strength classes have been constructed, and the results for a total of 810 specimens have been obtained. A comparative investigation was performed on two conditions of with and without considering the cement strength class as an input parameter in developed ANN-I and ANN-II models in order to obtain the optimum state.The comparison of the proposed idealized ANN model with two other existing models indicates good precision and accuracy of the developed ANN model in predicting the compressive strength of the mortar and the deficiency of these existing models in situations where cement strength class is present as an input parameter.