Non-destructive evaluation of the bond between a concrete added repair layer with variable thickness and a substrate layer using ANN

Abstract

The paper concerns the evaluation of the bond between a concrete added repair layer with variable thickness and a substrate layer, based on parameters assessed using non-destructive methods and also neural networks (ANN). For this purpose, two two-layered model elements were tested in a laboratory. For the evaluation of pull-off adhesion, five non-destructive testing (NDT) methods were proposed: the optical method used on the surface of the substrate layer, three acoustic methods used on the surface of the added repair layer and also the precise surveying method used on both the surface of the substrate layer and the added repair layer. Values of the pull-off adhesion obtained using the pull-off method in the same testing areas as using NDT methods were used for learning the ANN. It has been proved that such a method of pull-off adhesion evaluation is possible.