The HFRSET-based Detection Methods Of Broken Rotor Bar Fault In Induction Motors

The sensitive/reliable detection of broken rotor bar incipient fault in inductionmotors has been of profound importance in improving system safety, reducing outagetime and repair costs. Unfortunately, the current detection methods seriously suffer thefollowing problem: they are all puzzled by the limit of Frequency Resolution andMeasurement Time derived from Fourier Transform, on which they are generallybased, and thus they always lose their sensitivity/reliability in much reality whereunstable interference such as load fluctuation and random noise is inevitable. This hasbeen the key problem in the detection of broken rotor bar incipient fault in inductionmotors. Thus, it is a new challenge for the research on the detection of broken rotorbar fault in induction motors. This dissertation just addresses itself to thechallenge/trouble.In order to solve it, this proposal has deeply researched on the High FrequencyResolution Spectrum Estimation Technique, and smoothly combine them to detectbroken rotor bar incipient fault in induction motors, sensitively and reliably. Minimumvariance distortion response (MVDR) and its improved algorithm, multiple signalclassification (MUSIC) and its improved algorithm, which were the typical spectrumestimation techniques with high frequency resolution, had been tested with simulatedstator current signal of an induction motor with rotor bar fault. It shows that even witha short-time measurement data, they are both capable of correctly identifying thefrequencies of the broken rotor bar fault characteristic components, although theycan’t be used to handle with the amplitudes and initial phases of those components.Then the amplitudes and initial phases of the frequency components were estimated bythe least square method in extend Prony method. In addition, the overall performancecomparison between MVDR and MMVDR, along with that between MUSIC andRoot-MUSIC, had been made to pick out the superior for detecting broken rotor barfault in induction motors. Thereby formed a new method which were combinedMMVDR or Root-MUSIC with Prony algorithm to detect broken rotor bar fault ininduction motor.The related experiment on an induction motor is conducted and the resultsdemonstrate that the methods which were combined MMVDR or Root-MUSIC withProny algorithm to detect broken rotor bar fault in induction motor were effective evenwith short-time sample, as makes it a promising choice for broken rotor bar fault ininduction motors operating with fluctuant load or under severe interference. It can beexpected that the sensitive/reliable detection of broken rotor bar incipient fault ininduction motors will really become true through the above work, which is full ofpractical value and theoretical significance.