Prediction of compressive strength of lightweight concrete made with partially replaced cement by animal bone ash using artificial neural network

With the aid of an artificial neural network (ANN), an attempt has been made to build a model that can predict concrete's compressive strength. The data used to analyse and construct the model was generated from the laboratory experiment whereby concrete samples were saturated at 28 days. The compressive strength was taken using the Instron machine. The samples mix are divided into two, (1) normal conventional concrete, (2) lightweight concrete made with 100% pumice replacement with coarse aggregate and partial replacement of cement with animal bone ash (ABA). The use of an artificial neural network offers a non-destructive method of predicting concrete compressive strength. ANN is created in MATLAB using various NN and training functions and an objective criterion to produce the best model for compressive strength prediction. A gradient with magnitude weight bias learning and a mean square error (MSE) performance function was chosen for model training. The simulation results show that the ANN model correctly predicts the compressive strength of concrete with appropriate (MSE) and a correlation coefficient close to one.