Abstract
This research is concerning with the blow sound inspection method of the concrete using the neural network. It is generally difficult to diagnose the defective places of concrete by human beings. Then, using the neural network, the fault diagnosis method is proposed, in which, as a diagnosis standard, the spectrum data is obtained from the blow sound on the surface of concrete, and the normal state and the defective one are learned by the neural network. Using this diagnosis method, the defect of the concrete is found through the input of blow sound.