Visualization and Analysis Method of Defect Manifestation in Electromechanical Equipment

Abstract

This study focuses on the problem of diagnosing electromechanical equipment and aims to prevent its failures by timely detecting hidden signs of defects in diagnostic signals. This paper considers the possibility of improving systems whose equipment monitoring relies on measuring and analyzing the diagnostic signal of vibration or motor current. Fourier series decomposition for processing complex signals is not always effective because the contribution of harmonics reflecting the specific effect of the defect is less than that of non-specific harmonics and is comparable to the influence of noise. It has been proposed to apply the singular spectral analysis method for visualizing and analyzing the regularities of defect manifestations. It is reasonable to supplement the classical algorithm of this method by comparing the analyzed eigenvalue spectrum corresponding to the operating condition. Detection of hidden defects for the first time involves analyzing initial data projections in the directions of the singular basis that reflect deviations under the defect influence. Numerical and field experiments confirm the possibility of analyzing comparatively weak generations essential for equipment condition identification. The experiments demonstrate the opportunity for timely defect detection due to preprocessing when the probability of defect detection using the frequency method is close to zero. Thus, the approach to timely detection of equipment defects and making adequate decisions to manage its condition is justified. Doi: 10.28991/ESJ-2024-08-04-04 Full Text: PDF

Keywords

• diagnostic signal
• eigenvalue spectrum
• electromechanical equipment
• singular spectral analysis (ssa)
• visualization of defect manifestation.