A New Portable, Self-Powered, and Wireless Instrument for the Early Detection of Broken Rotor Bars in Induction Motors

A smart wireless sensor for the diagnosis of broken bars in induction motors

“…Such a high sampling rate is mandatory for detecting other kinds of faults like dynamic eccentricity for which features appear around 1 kHz in the spectrum. Similar peaks can appear in the spectrum of a healthy motor when it is operated with an oscillating load torque [7]. The motor has been driven using an inverter with scalar control.…”

Analysis of fault signatures for the diagnosis of induction motors fed by voltage source inverters using ANOVA and additive models

“…All these influences complicate the utilization of frequency analysis methods. VSI-fed motor faults have been analyzed and initial results are given in literature [21][22][23][24][25][26][27] but further investigation is still required [28].…”

Condition monitoring of electrical machines using low computing power devices

“…The mains component can be eliminated with the use of notch filters, [2], [3]. [12], spectral substraction [13], or, alternatively, the use of time-domain methods, like the dynamic time warping (DTW) [14], or the zero crossings times (ZCT) [15], [16] methods. • Spectral resolution must be high enough to properly resolve and identify the fault harmonics among a miriad of spectral lines that appear in the current's spectrum [9], even in the case of a healthy machine, due to the machine internal magnetic field [17], or to mechanical load oscillations [18].…”

“…Low sampling rates can be used if the current is smoothed with anti-aliasing filters before being sampled, as in [1], [2], [6], [16], [28], [29]. Nevertheless, a growing trend consists in oversampling the current signal at a very high sampling rate, and applying after a low pass digital filter.…”

“…Modern hardware can operate at sampling rates above 100 kHz [8], [27], [30], and has the computing power to perform real time digital filtering, so that high sampling rates are being increasingly used: 20 kHz in [31], 25 kHz in [32], 50 kHz in [33], and 100 kHz in [34]. • The diagnostic procedure must be simple, fast, and with low memory requirements, to be executed on line by embedded devices operating in real time [2], [11] such as DSPs [1], [6], [22], FPGAs [28], [35], portable devices [16], [36], small wireless sensors [33], [37], low cost processors [38], or energy-constrained sensor modules [39]. A growing trend is to implement the diagnostic algorithms in the same device that controls the power electronics in VSDs [40], [41].…”