The ANN Simulation of Stress Wave Based NDT on Concrete Structures

Abstract

As infrastructures are being deteriorated, non-destructive evaluation of existing structures and construction quality control are increasingly demanded. This paper is to study flaw detection technique of concrete members using Impact Echo method, which is one of the non-destructive test methods. This technique is based on stress wave propagation. The concrete members of the flaw detection consist of 5 types; one group for the thickness presumption, one group for the position of the aperture, one group for the position of the steel bar, one group that the steel plate is buried in it, and one group that PVC tube is buried in it. The Artificial Neural Network Technology was used for training of the thickness presumption of the concrete members. Experiments were performed using five types of concrete members that have simulated damages at known locations for tests and analysis. Test results are evaluated and summarized in this paper.