Predicting the significant characteristics of concrete containing palm oil fuel ash

Palm Oil Fuel Ash (POFA) is used as a supplementary cementitious material in concrete. Using different percentages of POFA leads to a non-linear variation among the characteristics of concrete. This study aims at developing an empirical model to predict the compressive strength of concrete using POFA as a cement replacement material and other properties of the concrete such as the slump and modulus of elasticity using an artificial neural network. Mixtures of concrete were selected with water-to -binder ratios of 0.50, 0.55 and 0.60, and 10%, 20%, 30% and 40% of the cement content was POFA. The 28-day compressive strength was tested, and the experimental results show that 0%–20% of POFA inclusion in the concrete mixtures has the most positive effects on the compressive strength. Then, a three-layer feed forward-back propagation ANN model with three inputs and three outputs was developed. Finally, the best architecture for the model was trained, tested and validated.