ANN based prediction of Bond and Impact Strength of Light Weight Self Consolidating Concrete with coconut shell

Abstract

In this experimental investigation, lightweight self-consolidating concrete (LWSCC) was developed with coconut shell as coarse aggregate. The effect of coconut shell aggregate (CSA) on bond strength and impact strength of Rice Husk Ash (RHA) based binary blended and RHA + Silica fume (SF) based ternary blended Self consolidating concrete (SCC) were determined. The bond strength was determined through pull-out test and the impact strength was calculated using falling weight test. The concrete mix was developed with the total powder content of 450 kg/m3. The coarse aggregate content was replaced by CSA in the gradation of 0%, 25%, 50%, 75% and 100% in the designated SCC. The investigation revealed that the bond and impact strength of CSA based LWSCC were comparable to current code practice and other lightweight concretes. The experimental data obtained was used to develop an ANN model for predicting the strength characteristics of fresh or hardened concrete. The high regression values obtained during training the neural network models reveals high accuracy and were predicting the strength characteristics very similar to the experimental results.