Hilbert Transform, an Effective Replacement of Park's Vector Modulus for the Detection of Rotor Faults

Abstract

The rotor-associated faults in electrical machines make the biggest most proportion of overall faults. Most of these faults particularly broken rotor bars, modulate all three phases of the stator in a symmetrical way. This fact makes Park's vector an effective tool to detect them at the incipient stage. As compared to other diagnostic algorithms, Park's vector is gaining more popularity for its noninvasive nature and comparatively simple mathematical model and a wide range of usage. In this paper, the broken rotor bars of a three-phase squirrel cage induction motor is detected using Park's vector modulus and Hilbert transform. It is shown that the Hilbert transform can be used to get the same kind of results as of Park's vector modulus but with a considerable decrease in data size required for analysis. The fault is first simulated using finite-element-based software for the study of its impact on various performance parameters and then Park's vector modulus and Hilbert transform are used to compare their required number of sensors and data size.