Research on Improving Seals to Suppress Vibration of Rotary Machines

Abstract

Based on the analysis of vibration signals modelled on a complex rotor system's impeller, such as an aviation gas turbine engine operating under steady-state conditions, this study demonstrates the influence of methodological factors – such as discrepancies in the natural frequencies of blade models, the sample length of the time series for the vibration process, and the sampling frequency – on the Hurst exponent value. The R/S analysis, or rescaled range method, proposed by H. Hurst for studying random processes, has provided a new perspective on the theory of random processes. This method has also become a highly effective and widely applied technical approach for examining a wide range of processes, from telecommunications and bioelectric signal analysis to solving functional diagnostic tasks for complex spatial objects. The findings of this study may be implemented within functional diagnostic systems for complex rotor systems to detect and predict the behaviour of damage in their rotating components.