Self-organizing map neural network for transient signal classification in mechanical diagnostics

Acoustic Emissions (AE), generated by the formation and growth of micro-cracks in metal components, provide us with a promising mechanical fault detection technique in monitoring complex-shaped components in helicopters and aircraft. A major challenge for an AE-based fault detection algorithm is to distinguish crack related AE signals from other interfering transient signals, such as fretting related AE signals and electromagnetic transients. In this paper, we presents a classifier, which makes its decision based on the features extracted from joint time-frequency distribution data by Self-Organizing Map (SOM) neural network. In-flight data are used to test the performance of this classification system, with promising results.