Bearing Fault Detection and Quantification for Flight Control Electro-mechanical Actuators

This study aimed at investigating fault detection and quantification methods for flight control electro-mechanical actuators (EMAs). The fault quantification work is related to spall quantification for ball screws and ball bearings. Several quantification methods originally developed for industrial applications are intensively reviewed. Such methods are inadequate for EMAs for which a wide variety of transient fault types occurs. A new vibration-based quantification method has been developed to estimate spall size for ball screws and ball bearings within EMAs. The principle is based on locating geometric features on the vibrational jerk rather than direct acceleration measurements. The jerk is numerically derived from acceleration data using a Savitzky-Golay differentiator, which provides signal enhancement to accurately identify spall-related geometric features and their physical size. Experimental work involves fault injected data sets collected from three dissimilar EMA test stands, two provided by the German Aerospace Center (DLR) and one by the NASA Electro-Mechanical Actuator test stand.