Improvement of fault diagnosis reliability and depth is based on the application of fault criteria system, which is formed by combining the various parameters of diagnostic signals.
Therefore, the extension of the scope of signal diagnostic parameters and fault symptoms is important for diagnosis. The paper discusses the possibility of applying probabilistic characteristics both numerical and functional to estimate the parameters of vibration processes that characterize the defects and faults of the various mechanisms. Studies of vibration parameters when changing the condition of rolling bearings within various units and mechanisms showed that the probability density of the instantane-ous values oftbe vibroacoustic signal, and in particular, the acceleration also changes its parameters wben the bearing condition changing. Similar results were obtained when analyzing the probabilistic characteristics of the instantaneous values of vibroacoustic signals received from different assamblies of reciprocating compressors. We found that the change in the condition of such reciprocating compressor parts as discharge and suction valve, details of piston-cylinder units, slide-crank and cranking mechanisms, the crank-shaft bearings, the probability curves of the instantaneous vallies of the vibration signal significantly change. The latter fact is clearly visible when analyzing the parameters of the approximated distribution functions and probability density. For the first time in vibroacoustic diagnostics it was proposed to use such probabilistic characteristic of random processes as a characteristic function to analyze the properties of distribution function curve and probability curve.
The studies have shown that the area under tbe curve of the modulus of the characteristic function of the instantaneous value of vibroacoustic signals of different components and mechanisms may be used as a diagnostic parameter. Formation of the diagnostic features vector based on the parameter has allowed to propose a new diagnostic feature of fault invariant to machine design and the limit values ot vibroacollstic signal. Based on revealed regularity there have been developed a method and techniques of vibroacoustic diagnosis of various units of machines and mechanisms, as well as the vibroacollstic signal processing algorithm. Application of the deveIoped diagnosis method and diagnostic characteristics based on the characteristic function of vibroacoustic signal alows to increase the depth and the reliability of diagnosis whiIe simplifying diagnosis algorithms.
V. N. Kostyukov, А. Р. Nаumenkо, S. N. Boychenko (DYNAMICS Scientific & Ргоduсtion Center, Ltd., Omsk, Russia); 1. S. Kudr'yavtseva (Omsk State Тесhniсаl University, Omsk, Russia) Formation of diagnostic feature vector based on characteristic function of vibroacoustic signal // Нефтепереработка и нефтехимия. - 2016. - № 8. - 51-56
Download the publication
You can order the translation by e-mail: email@example.com