The measurement of acoustic emission data in experiments reveals informative details about the tribological contact. The required recording rate for conclusive datasets ranges up to several megahertz. Typically, this results in very large datasets for long-term measurements. This in return has the consequence, that acoustic emissions are mostly acquired at predefined cyclic time intervals, which leads to many blind spots. The following work shows methods for effective postprocessing and a feature based data acquisition method. Additionally, a two stage wear mechanism for bearings was found by the described method and could be substantiated by a numerical simulation.
