THE USE OF A WAVELET TRANSFORMATION FOR REMOVAL OF THE NOISE COMPONENT FROM THE VIBRO SIGNAL

Abstract

The article investigates the use of wavelets to remove noise from the measuring vibration signal. It is determined that wavelets are well adapted for signal analysis, for which the principle of causality is important: wavelets preserve the direction of time and do not create parasitic interference between the past and the future. Criteria for selecting an analytical wavelet have been developed, depending on what information should be extracted from the signal and the need to more fully identify and emphasize certain properties of the analyzed signal. It is proposed to use Daubechies wavelets to process the vibration signal data. The simulation of vibration signal filtering from noise with the normal distribution law is performed in the MATCAD package. It is proved that the method of wavelet transform allows to solve the problem of filtering the vibration signal from noise when processing vibration signals obtained by autonomous recording devices in conditions of increased interference from the environment. The obtained results evidence to the prospects of the developed method and its advantages in comparison with the hardware solution of the filtering problem.