Diagnosis indicator for blade-casing rubbing faults based on multi-harmonic phases

Abstract

Blade-casing rubbing is a typical fault of aero-engines. Due to its scattered features and primarily caused by other faults, diagnosing it remains a challenge. This paper proposes a diagnosis indicator based on multi-harmonic phases for real-time monitoring of rubbing faults in aero-engines. Firstly, through the theoretical derivation, we find that the multi-harmonic phases can reflect the occurrence of rubbing faults. Secondly, a rubbing fault diagnosis indicator, PKLM, is constructed by weighted incorporation of multi-harmonic phases, with Kullback–Leibler divergence employed to amplify the differences between the fault and normal conditions. An optimal harmonic number criterion, S, is established for constructing the indicator, along with a threshold, PKLH, to determine the system’s condition. Thirdly, the measured harmonic phases are processed to serve as effective inputs for the indicator. Finally, model simulations and experimental results confirm that the proposed indicator can isolate interference from other rotor faults and effectively quantitatively diagnose aero-engine rubbing faults.