INTELLIGENT TECHNOLOGY OF WAVELET ANALYSIS OF VIBRATION SIGNALS

Abstract

During solution of engineering problems of machinery dynamics a need of revealing the harmonic components often arises in the narrow timing gate. This requires the use of wavelet-transformation oscillation methods and introduction of intelligent systems to hardware and software used in the experiment. The wavelet is considered as a short in time signal functional window, which has its internal structure in the form of a fading wavelike burst, and it is characterized by a scale of display of certain events in the field of the signal frequency spectrum, as well as and by time axis shifts. Complex-functioned continuous functions of real arguments (Daubechies wavelets, Gaussian wavelets, MHat-wavelets), complex-valued functions of real arguments (Morlet and Paul wavelets), as well as real discrete functions (HААRT- and FHat-wavelets) are used as wavelet functions. The wavelet analysis method of vibration signals is disclosed at acoustic diagnostics of machines and mechanisms. Digital implementation of discrete indications of wavelets with the subsequent visualization of results in the form of scalotons is the mathematical basis of the algorithm for procession of vibration signals. It has been suggested that engineering analysis and reconstruction of signals should be implemented by means of directed and reverse continuous wavelet conversions, which are discrete by arguments. The structural and functional scheme of the multichannel system of the intelligent wavelet analysis of vibration signals in machines has been considered. The intelligent system for study of vibration signals makes it possible to form the totality of photographic parameters, when scalotons are calculated by wavelet functions. An example of experimental implementation of the wavelet conversion method of vibration signals parameters is shown. Results of scalotons calculation are shown, when MHat-wavelet and DOG-wavelet are used.