The diagnosis method of concrete structures using non-destructive tests based on stress waves and artificial neural network

As infrastructures are being deteriorated, the necessity of non-destructive evaluation in existing structures and new building construction for the quality control are increasingly demanded. This paper is to study the flaw detection techniques of concrete members using an impact echo method, which is one of the nondestructive test methods. This technique is based on stress wave propagation. The flaw detections of concrete members consist of 4 types in this study; one group for the thickness prediction, one group for the position of voids, one group for the position of the reinforcement bars, and the other group for the position of PVC tubes buried in the specimens. The artificial neural network technology was used for training of the thickness prediction for the concrete members. Experiments were performed using four types of flaws simulated in the concrete specimens at known locations for tests and analysis. Test results of flaws are evaluated and presented in this paper.