Prediction of Compressive Strength of Aerated Lightweight Aggregate Concrete by Artificial Neural Network

This paper presents artificial neural network techniques for predicting the compressive strength of Aerated Lightweight Aggregate Concrete (ALAC) based on the effects of the concrete mix parameters. The compressive strength of sixty different concretes with densities ranging from 551 to 1948 kg/m3 was used and trained. The primary mix design variables studied included amount of cement, water, coarse aggregate, fine aggregate, surfactant, the volume percentage of air in the matrix (A/M), and the volume percentage of matrix of the total mix (M/T). The training and testing results indicate that the model explains 0.984 and 0.979 of the variability in compressive strength for the single aggregate used in the study, respectively.