Classifying helicopter gearbox faults using a normalized energymetric

Abstract

A normalized energy metric is used to classify seeded faults of theOH-58A main transmission. This gearbox comprises a two-stagetransmission with an overall reduction of 17.44:1. Loaded gearbox test runsare used to evaluate the sensitivity of a non-stationary fault metric forearly fault detection and classification. The non-stationary fault metricconsists of a simple normalized energy index developed to account for aredistribution of sideband energy of the dominant mesh frequency and itsharmonics in the presence of actual gearbox faults. This index is used toqualitatively assess the presence, type and location of gearbox faults. Inthis work, elements of the normalized energy metric are assembled into afeature vector to serve as input into a self-organizing Kohonen neural networkclassifier. This classifier maps vibration features onto a two-dimensionalgrid. A feedforward back propagation neural network is then used to classifydifferent faults according to how they cluster on the two-dimensionalself-organizing map. Gearbox faults of OH-58A main transmission considered inthis study include sun gear spalling and spiral bevel gear scoring. Resultsfrom the classification suggest that the normalized energy metric isreasonably robust against false alarms for certain geartrain faults.