Abstract & Overview
Compressive strength is an important aspect in the design of concrete structures. Excepting for laboratory casting and testing of concrete samples, no reliable method is available to evaluate the compressive strength of in-situ concrete precisely. The objective of the paper was to develop a different methodology for the determination of concrete compressive strength based on measured acceleration using Artificial Neural Network (ANN). The surface of a cube was struck with an impactor and induced acceleration reading at six different locations on it were recorded. Strength of concrete was determined by testing. Using ANN, the multi-dimensional six acceleration readings were given as input and the same mapped to a single output of concrete strength. The range of acceleration values varied from 0.7 m/s2 to 136.4 m/s2 and compressive strength from 16 to 43.56 MPa. The values pertaining to 54 cubes were used to train the network. The correlation between predicted and measured strength was found to be good