A Novel Data Fusion Method for Incipient Fault Detection in TRU of Aircraft Electrical System

摘要

Incipient fault detection in the Transformer Rectifier Unit (TRU) can prompt early preparedness for replacement of Line Replaceable Units (LRUs) thereby reducing the downtime of the aircraft. This paper describes a novel incipient fault detection and fault localization method of TRU in the context of an aircraft electrical system. Results from Failure Mode and Effects Analysis (FMEA) study identified diode failure as the primary reason for failure of TRU. The incipient fault detection of TRU is done by applying data fusion algorithm. The algorithm uses Hilbert transform as one of the computing tools to derive Health Index Parameter (HIP). Fast Fourier Transforms (FFT) and wavelet transforms were used for fault localization of the component within the TRU. The developed incipient fault detection method is validated on a TRU test rig. Though the test rig is configured on 50Hz supply the results can be read across on 400Hz supply. A scheme of communicating the fault status integrating through 1553B network is also addressed.