New technique of nondestructive assessment of concrete strength using artificial intelligence

A new technique of nondestructively assessing the compression strength of concrete, which employs artificial neural networks, is presented. A data set based on results of testing concretes (with a 28-day strength of 24-105 MPa) by different nondestructive methods was used to train and test an artificial neural network. The methodology of neural identification of the strength is described. The obtained results, including those of the practical verification of the technique, are reported.