Predicting the compressive strength of a quaternary blend concrete using Bayesian regularized neural network

Concrete produced with ordinary Portland cement (OPC) along insertion of supplementary materials increases the level nonlinearity. Due to this increased non-linearity and difficulty in modeling numerically, focus has on exploration computational intelligent models like artificial neural network (ANN) estimate different concrete properties. In study, a quaternary blend was developed OPC, fly ash (FA), metakaolin (MK) rice husk (RHA). The experimental data were further used training proposed ANN approximate its compressive strength. trained optimized using three regularization algorithms; scaled conjugate gradient “trainsc” (SCG), Levenberg–Marquardt “trainlm” (LM) Bayesian regularized “trainbr” (BR) algorithms. percent proportion FA, MK RHA making up blends curing days are five features as input variables, while strength each individual mixture is output variable (target). It found out that function performed best highest correlation coefficient, lowest MAE, MSE RMSE. results obtained from approach show significant improvement observations.