ESTIMATION OF ACOUSTIC-EMISSION SOURCE WAVE-FORM OF FRACTURE USING A NEURAL-NETWORK

The applicability of a neural network to acoustic emission (AE) is pre sented. It is shown that the shape of the simulated source waveform us ing piezoelectric ceramics is steplike, similar to that of mode I crac k extension, and its vise-time can be varied by the resonance frequenc y in the thickness direction. The results imply that the simulated sou rce can provide learning waveforms for the network. Actual AE waveform s were also acquired by conducting a tensile test of a chevron-notched graphite specimen. It was demonstrated that the appropriate source wa veform associated with mode I crack extension was successfully determi ned by the network taught with simulated waveforms.