Early fault diagnostics of aero-engine rotor bearing using AdaESPGL algorithm with optimal multiplier

This study focuses on the practical engineering solutions for aero-engine bearing diagnosis. As one of the leading factors for in-flight shutdown and unplanned engine replacement, bearing fault diagnosis and prediction have been extensively investigated. Until today, there are still unanswered questions on aero-engine fault diagnostics. Both theoretical and numerical, as well as empirical studies, are needed to address those questions. Applying naturally induced aero-engine rotor bearing failure data, this paper evaluates the sparse decomposition (SD)-based algorithm for bearing diagnostics. It proposes optimizing the multiplier for regularization parameter estimation to improve the early bearing fault diag nosis accuracy. The advantage of the proposed algorithm improvement is then validated. This paper concludes with recommendations for practical aero-engine bearing fault diag nostics development methodologies and procedures.