Prediction of compressive strength of “green” concrete using artificial neural networks

With its growing emphasis on sustainability, the construction industry is more interested in applying environmentally friendly concrete, also known as “gre en” concrete, in its construction projects. Among other benefits, concrete made with alternative or recycled waste material can reduce pollution and energy use, as well as lower the cost of concrete production. However, the impacts of these alternative mat erials on concrete properties have not been fully understood, which limits the wide applications of “green” concrete in practice. This study investigates the application of Artificial Neural Networks (ANN) to predict the compressive strength (CS) of concr ete made with alternative materials such as fly ash, Haydite lightweight aggregate and Portland limestone cement. A feed - forward Multilayer Perceptron (MLP) model was applied for this purpose. To determine the accuracy and flexibility of this approach, two different input methods (relative and numerical) were tested on the generated ANN models. The results showed that concrete made of Portland limestone cement had slightly better CS than concrete made of Portland cement. Generally, both input methods provid ed adequate accuracy to predict CS. It was also observed that a proper MLP model with one hidden layer and sufficient neurons (depending on the input variables and type of cement) could effectively predict the CS of “green” concrete.