Increasing the sensitivity and expanding the functionality of bipolar analysis of rotary machine vibrations

Abstract

Vibration-acoustic diagnostics of rotary machine defects is the most effective method for non-destructive testing of their technical condition. Practice shows that its successful use largely depends on a set of available methods for analyzing vibration processes. Gears are the most common and heavy-duty components that largely determine the overall vibratory condition of a machine. There is quite a variety of methods for vibration-based diagnostics of gear defects. They include an interesting method of bipolar analysis which consists in separate analysis of the positive and negative parts of a vibration signal with subsequent formation of the diagnostic indicator in the form of differences, ratios, etc. The method is aimed at assessing the quality of gearbox assembly by the position of the tooth contact pattern. Limited area of use and low sensitivity is a disadvantage of this method. The paper shows that the use of broadband vibration maxima in the application of bipolar analysis in vibration-based diagnostics of rotary machine defects significantly increases its efficiency. Using the example of wear of tooth flanks in the sun gear satellite gears pair and the value of the backlash in the differential gearbox of a turboprop engine, it was found to be ensured by increasing the sensitivity of the method and expanding its functionality. In our analysis, we used statistics of gearboxes with different degrees of wear of the tooth flanks and repaired gearboxes with different backlashes. In this case, a wide set of well-known diagnostic indicators can be used: intensity of n-dimensional vectors of informative series, parameters of individual harmonics, amplitude modulation depth, probabilistic characteristics in selected frequency bands, dimensionless discriminants, cepstra, etc. The paper presents some examples of these methods.