Localization of AE Source in Plates Using ANN Approach: An Experimental Investigation

Structures are prone to damage, and detecting them early is extremely important for taking corrective measures. It is understood from the literature that the location of damage acts as an acoustic emission (AE) source. Acoustic signals emitted from the source are captured by AE sensors and then analyzed to detect the source. Effective localization of AE source using conventional methods of signal analysis depends on detecting the actual arrival time of the AE waves at the sensors, use of appropriate material property, concept of relevant signal processing techniques, etc. In this context, artificial neural network (ANN) technique is very promising. While the effectiveness of ANN depends on data of good quality and quantity, use of an established ANN model is very simple. In light of the above, the present study makes an effort to develop an ANN model for prediction of AE source location in a plate. For developing the ANN model, adequate data are collected through extensive experimentation with aluminum plate in the laboratory setup, where pencil lead break (PLB) is used as an AE source and a single AE sensor is utilized. Relevant signal features are extracted from each signal and used as the inputs to the model whereas distance between the source and the sensor is taken as output. The trained model is then tested with another set of experimental data. The results of the AE source locations, predicted by the model, are found to be very encouraging when compared with the actual location of the source.