Method of separation of failure symptoms identified during run-up or run-down conditions of machine operation

Investigation of the rotating machinery in transient conditions have a lot of advantages. Analysis of signals recorder during these investigation leads to the time-frequency representation. Effects of the signal analysis can allow not only to identify signal components but first of all their exploitation in time. Identified signal components are symptoms of phenomena related and not related to the changes of conditions of machine operation. Signal analysis (applied in investigations in varying conditions of machine operation) does not make it possible to separate the above mentioned symptoms. It has very often an influence on the correct determination of machine state. The resolution of this problem can be application the RLS analysis which consists in separation of identified symptoms on the time frequency characteristics. The basis of this analysis is an assumption that recorded signal is result of resonance and periodic excitations. The paper deals with investigations of rotating machinery in run up or run down conditions. The presented results are effects of applications of two kinds of signal analysis: method based in the Fourier transform and method based on the wavelt transform. The letter analysis was not applied in this kind of investigations. The effects of both analysis are time frequency characteristic which are input data to the application of the RLS method.