Vibrational Analysis Techniques for Rolling Bearings at low rotational speeds =

Bearing vibration analysis plays a crucial role in the condition monitoring and predictive maintenance of rotating machinery. However, when rotating industrial equipment operates at low speeds, the vibration signal is affected with problems, which makes it more challenging to extrapolate information regarding the health status of the components. This thesis, written in collaboration with Remote Diagnostic Center at SKF Industrie S.p.a., aims to explore advanced vibration monitoring techniques for rolling bearings at low speeds in order to improve fault diagnosis. Vibration diagnostics makes use of the signal from accelerometers placed on the housing close to the bearings. Successively, the acquired signal is processed to obtain information about the health of the bearings. This process involves extracting parameters known as Condition Indicators from the signal to identify the possibility of defect, followed by spectrum analysis to detect the type and severity of damage. In this thesis advanced techniques such as Spectral Amplitude Demodulation, Autogram and Spectral Correlation are implemented. This is done to optimize the spectrum analysis, since at low rotational speeds, the signal is much weaker and the frequencies of interest are very small and very close together. These methodologies enable the spectrum to be extracted more effectively from the enhanced envelope of the raw signal. In addition to these, another advanced methodology developed in-house by SKF, the Repetitive Fourier Transform will be compared with the classic Envelope Spectrum, already used by SKF for the diagnosis of rotating machinery. Tests will be conducted using sample bearings, artificially damaged, under various speed and load conditions. These tests will be carried out at the Global Life Testing Centre in Airasca, Italy. The specimen bearings, series SFK 7209 BEP angular contact ball bearings, will be artificially damaged. In this thesis damages will only be carried out on the inner and outer raceway with three different intensity levels, while varying rotational speeds between 500 rpm and 70 rpm. The workbench, an SKF R2 RCF test rig, puts a variable axial load on the bearings. Furthermore, to avoid signal interference, only one damaged bearing was tested per time. Moreover special care was taken during mounting to fit the samples damaged on the outer ring with the damage placed in proximity to the sensor in order to strengthen the faulty signal as much as possible. Condition indicators will then be extrapolated from the raw signals from the tests. These will then be cross-compared to check their effectiveness in detecting any anomalies/damage. The advanced techniques, as well as the SKF's internal techniques, will be used to produce different types of envelope spectra for a more in-depth analysis of the type and severity of the damages. To verify the efficiency in facilitating diagnostics, will be used a dummy parameter indicated as the sum of the first three harmonics of the defect frequency divided by the RMS value of the enhanced spectra obtained. This parameter is to be interpreted as a value indicating how much more pronounced and visible the defect frequency peaks are. Finally, the results collected will be statistically analyzed and discussed to determine which of these methods is most suitable to use under the conditions specified above, while also considering their feasibility and highlighting appropriate differences with real industrial applications.