[106] Development of a bearing fault detection and diagnosis suite based on multi-sensing techniques for aero engines

Rolling element bearings provide vital load-carrying roles in many rotational applications. In aero engines these include supports for the main shafts and the main gearbox. While advances in material manufacturing techniques for producing steels of very high purity mean that unexpected bearing failures are very rare, when they do occur the consequences are severe. Thus, there is a growing desire for prognostic health monitoring (for example remaining lifetime) of these vital components so that preventative action can be taken to eliminate main shaft bearing failure.
This project has been exploring the uses of multiple sensing techniques that can survive the harsh conditions encountered in the bearing chambers of aero engines and provide information on the health of operating bearings. Sensors chosen, including accelerometers (to measure vibration), thermocouples (bearing temperatures), strain gauges (force), inductive and eddy current technologies (shaft and cage/ball speeds), to monitor bearing health via the detection and diagnosis of defects in bearings under a wide range of conditions are discussed.
The sensors have been implemented on a bearing test-rig and artificial defects of different levels/sizes have been implemented on the components of test bearings (inner and outer races, as well as rolling elements). Additionally, a shaker is connected to the test-rig to add noise either at a specific frequency or broadband (white noise) during the tests to simulate noisy aero engine environments.
The data obtained will be used to train bearing monitoring algorithms and to develop fault detection and diagnosis strategies. Detailed results from the subscale testing and the validation of bearing health monitoring strategies will be presented.