Modeling, Design and Testing Techniques for New Generation Industrial Rotors

This dissertation explores relevant topics regarding the application of machine fault diagnosis in the design, monitoring, and diagnosing of industrial rotating machinery. Bearings, fundamental components in rotating systems influencing the dynamics of the entire system, are affected by roundness errors developed during production and assembly and may suffer damage during operation, causing unwanted vibration with a detrimental effect on the efficiency and quality of production. In the context of Industry 4.0, sensor-equipped rolling bearings can be used as a sentry for the whole rotor system, providing both internal bearing damage detection and external rotor system anomalies monitoring. This dissertation focuses upon the installation, sensorization, and experimental activity on a test rig for the monitoring of medium-sized industrial bearings while operating in industrial conditions, available at the Laboratory of Mechanics of Politecnico di Torino. The state-of-the-art of modeling techniques for rolling bearings with localized defects was also examined. An analytical numerical model of the rotor-bearing system was developed, validated through some experimental data acquired on the test rig, able to identify vibrational features from defects and to perform component- and system-level monitoring activities within its vibrational response. A dataset was created for condition monitoring of medium-size spherical roller bearings with various localized defects. Noise reduction in experimental vibration signals, particularly using the Discrete Wavelet Transform (DWT), was also investigated using an experimental literature dataset. A screening of DWT parameters was conducted for the appropriate parameters selection for the identification of localized defects in bearings, enhancing previous literature studies. Furthermore, the effect of the geometry of the bearing installation seat on the roundness error of bearing inner ring was also analysed. Extensive experimental activity, performed in the ARotor Lab at Aalto University, validated the roundness error stacking methodology to evaluate the roundness error of mounted inner ring raceways considering the roundness error of the tapered shaft and the thickness variation of the inner ring. The existence of an optimal orientation between the components that minimizes the roundness error of the raceways in the assembly was highlighted. The study further explored the application of 3D conical grinding to the tapered installation seats of a large rotor to compensate for roundness error. A significant reduction in the roundness errors of installed inner ring raceways and rotor's subcritical vibration was experimentally demonstrated, also compared to previously published methodologies based on cylindrical grinding and the use of adapter sleeves. Results showed that unmounted bearing inner ring roundness is irrelevant to the final raceway roundness, and that bearing manufacturers should focus on minimizing the thickness variation of the bearing inner ring.