کنفرانس بین المللی ماشین و محرکه های الکتریکی

صفحه اصلی / The 5th International Conference on Electrical Machines and Drives

Fuzzy Logic Based Severity Detection of Broken Rotor Bars Using Current and Vibration Analysis

نویسندگان :

Hamidreza Aghaei¹ Ebrahim Nikghalb² Omid Khanzadeh³

1- Electrical Specialist Electrical Machine Laboratory Electric Motors SBU MAPNA Generator Engineering and Manufacturing Co. (PARS) 2- Head of Electrical Machine Laboratory Electric Motors SBU MAPNA Generator Engineering and Manufacturing Co. (PARS) 3- Manager of Electric Motors SBU MAPNA Generator Engineering and Manufacturing Co. (PARS)

کلمات کلیدی :

induction motor،broken rotor bar fault،current and vibration analysis،fault severity index،fuzzy logic،slip estimation،windowed FFT

چکیده :

Broken rotor bars constitute a critical rotor fault mode in squirrel cage induction motors, whose severity quantification beyond mere detection directly informs maintenance planning and risk mitigation. A load-robust severity estimation framework is introduced, built upon motor current signature analysis while explicitly addressing operating-point dependence. First, instantaneous slip is estimated by jointly using stator-current and vibration measurements, enabling precise localization of sidebands, even under partial load where amplitudes diminish and may overlap the fundamental. Next, a windowed fast Fourier transform with leakage controlled windowing stabilizes downstream features. On these slip-aligned spectra, sideband amplitude features are extracted and normalized to operating point to form a load-independent severity index, which is mapped to actionable levels via fuzzy logic classification. Across multiple loads and 1–4 broken-bar conditions, the proposed index yields more consistent separation than a conventional amplitude baseline and improves severity classification accuracy by 18, 20.8, 3.5, and 7.8 percentage points for the four, three, two, and one broken bar cases, respectively, while slip alignment keeps the sideband-frequency error below 0.4% and windowing restores sideband visibility at low slip.