Abstract
Advanced analysis of motor currents during transient operation has proven to be a valuable source of information to reach a reliable assessment of the rotor condition. This novel approach may help to overcome some of the drawbacks of the classical Motor Current Signature Analysis (MCSA). One of the advantages of the approach is the robustness obtained due to its twofold perspective: 1) Qualitative identification of fault-related time-frequency patterns + 2) Computation of quantification indicators to determine the severity of the fault. Nonetheless, since the research in this area is still incipient, few transient-based fault severity indicators have been hitherto proposed in the literature. Moreover, most of them have been validated for specific startup conditions (the majority
for line-started motors). This paper presents a study about the variation of different wavelet-based fault severity indicators for different types of motor starting methods. An experimental testbed enabling the development of different types of startups with different rotor fault levels enables to obtain real startup currents that are used for the study. Results show the robustness of the indicators even in situations where the classical MCSA does not lead to correct results.
