An artificial neural network (ANN) approach was used to characterize v
ibration data for nondestructive evaluation purposes. Acoustic signatu
res were obtained from clamped-clamped metal beams of rectangular cros
s section. The beams were either intact, or had one (small) slot in th
em. The digitized data were used to train the ANN to predict future sa
mples of the measured time series given past and present samples. The
trained ANNs were used in two ways. In the first method, an ANN was tr
ained with vibration data from intact beams. Once the ANN could adequa
tely predict the training signal, vibration signals obtained from beam
s with slots were presented. Significant differences between predictio
n errors for the intact beam and beams with slots as shallow as 0.1 in
. were found. Furthermore, the resulting prediction errors gradually i
ncreased as the slots in the beams grow deeper, suggesting that this m
ethod is useful to estimate defect size. In the second method, the con
nection weights of the ANNs trained on vibration data from intact beam
s were compared to the corresponding weights of ANNs retrained on the
test data. Again, this approach was very sensitive, but no useful rela
tionship with the slot depth was found. (C) 1995 Acoustical Society of
America.