Cyclostationary analysis of vibration signals from electric motor – understanding of bi-frequency map

Abstract. Cyclostationary analysis is one of the most promising and powerful methods for analyzing vibration signals from rolling element bearings (REBs). Damage to the REB is one of the most common reasons for failure in the electric motor (EM). Thus, the application of cyclostationary analysis to vibration signals from EM is reasonable in order to detect faults. Unfortunately, it has been observed that analyzing vibration signals measured on bearings in electric motors is much more challenging due to the presence of additional cyclic components on the bi-frequency map that affect the analysis. These additional components may have various mechanical sources (for example, the influence of shaft misalignment) or electrical/electromagnetic origins (such as frequencies related to features of the power source, i.e., grid frequency, inverter keying, etc.). This paper is an introduction to bearing diagnostics in electric motors based on cyclostationary analysis in the bi-frequency domain. We present various cases to show the complexity of the map and conclude that there is a need to define a method for automatic identification and interpretation of sources, i.e., understanding of the bi-frequency map. As other cyclic sources are wideband and they have much stronger values on the map, simple band selection is difficult. Thus, new methods for further map processing and novel indicators to assess the significance of features that describe the fault based on cyclic spectral coherence (CSC) maps should be developed. However, it is stated that it requires a deep understanding of the CSC map. We believe that this paper brings some new knowledge to the context of using CSC for bearing diagnosis in electric motors.