AGMA 18FTM10 PDF

General Topic:

To run all components of an aero engine at optimal speeds a planetary gearbox between fan and turbine is required. This new aero engine design allows slow fan rotation together with high turbine speed. As a result, the efficiency and bypass ratio increase significantly. However, using a gearbox introduces additional failure modes such as gear wear, pitting and gear teeth cracks.

Usually, high attention is paid to pitting initiation since this failure mode leads to subsequent destructive failures. Therefore, the early detection of pitting using a feature-based condition monitoring system is recommended to avoid unplanned engine shutdown and expensive gear replacements.

This paper describes an experimental investigation on spur gears to characterize the pitting degradation process using monitoring features.

Specific Question or Relationship:

Previous investigations have revealed that pitting has an impact on the gear vibration behavior [27]. But is it possible to detect pitting at an early stage using acceleration sensors and acoustic emission (AE) sensors to avoid consequential damages and subsequent correction activities?

It is assumed that the pitting initiation process begins with the excitation of high frequency AE impulses above 50 kHz. Moreover, pitting progression causes increased sidebands in the gear mesh order spectrum. Therefore, the AE sensor technique with a higher bandwidth seems to be more suitable for early pitting detection.

The main goal of this experimental investigation is to understand, if an early pitting detection can be achieved using AE sensors combined with appropriate signal processing methods.

Method:

The authors investigated several gears manufactured with different pitch deviations and surface properties. The gears were pre-loaded in a back-to-back test rig and driven until distinctive pitting occurred. Multiple acceleration sensors and AE sensors are mounted on the stationary housing adjacent to the gear bearing shell.

The sensors measure the vibration characteristics of the intermeshing gears. The signals were then filtered, resampled and pattern recognition methods were applied to extract monitoring features [4]. These features allowed a conclusion to be made regarding the correlations between those features and the pitting degradation process of the gears.

Results:

For pitting detection, the AE sensors seem to be more suitable due to their higher bandwidth in comparison to the industrial acceleration sensors, for which the bandwidth ends around 50 kHz. The results show that a detection of pitting is possible several hours before complete gear failure. The test run results also depict the advantages and disadvantages of both the acceleration and the AE sensors.

Conclusion:

The results indicate that early detection of pitting damage is possible. Based on the results, the requirements for potential production sensors and the corresponding signal processing algorithms can be defined for the integration in the control monitoring system for future aero engines. Furthermore, the influence of manufacturing processes and manufacturing deviations on the pitting progression can be investigated. A correlation between different finish qualities and the last machining step must be assessed to identify the associated impact on surface integrity.