Özet:

Modern condition monitoring systems (CMS) collect and process enormous amount of data in order to provide the earliest and most dependable information of fault development within any of the machine components and their operation combined. According to numerous studies one of the most fault susceptible mechanical elements in rotating machinery are rolling element bearings. Although reliable techniques for their diagnostics are already proposed, the new investigation is needed. According to authors experience in many industrial applications the operators are obligated to simultaneously track hundreds of diagnostic estimates, such as signals energy, its peakedness or narrowband characteristics for localized faults. As mentioned, for a vibration-based CMS of single wind turbine there are nearly 150 of them. Therefore, the authors employ a fuzzy logic approach for integration of bearing diagnostic features. A new estimate that carry most relevant information about bearing condition is discussed. The reasoning is presented on simulated data that mimics real rotating machine.

Anahtar Kelimeler: