The present work concerns the development of a Lamb-wave-based imaging approach with the capacity to visually pinpoint structural damage, if any, in terms of the probability of damage occurrence at all spatial positions of the structure under inspection. To establish such probabilities, individual sensors of an active sensor network contributed their perceptions as to the damage occurrence near them using the signal feature time-of-flight (ToF) extracted from captured Lamb wave signals. All these perceptions were then fused by virtue of an image arithmetic algorithm. The prediction results were presented in an image where the location and size of all the damage instances in the structure became intuitional, rather than provided with definitive damage parameters. Such a probability-based imaging approach is by nature more consistent with the implication of prediction or estimation of damage than traditional identification endeavours. The effectiveness of the approach was experimentally demonstrated by predicting delamination in carbon-fibre-reinforced epoxy (CF/EP) laminates.
展开▼
