The Detection of Rotor Bar Faults in Induction Motors Using the Recursive Matrix Pencil Method

Alexander Shestakov, Olga Ibryaeva, Victoria Eremeeva, Vladimir Sinitsin
Abstract:

Rotor cage induction motors are widely used in the industry and their unpredicted shutdown can be very costly. Therefore, for safety and economic considerations, there is a need for identifying incipient faults. Among the induction motors faults, the rotor bar faults are among the most common failures. These generate sideband frequency components around the frequency of the power supply. The amplitudes of these sideband frequencies increase with the progression of the bar fault. Here we propose that the Recursive Matrix Pencil Method is able to track the growth of the sideband frequency amplitude excited by a rotor bar fault and show the results of applying the technique to numerically simulated data. Signal preprocessing consists in filtering the supply frequency and bandpass filtering with a passband frequency range specified by one of the sideband frequencies. The paper uses the modelling of the induction motor current signal so that the gradual development of a rotor bar fault can be simulated. It is shown that the Recursive Matrix Pencil Method gives a time advantage over the Classical Matrix Pencil Method and the possibility of practical application.

Keywords:
Matrix Pencil Method, Induction Motor, Rotor Bar Faults, Diagnostic, Current Signal Analysis
Download:
IMEKO-TC10-2023-015.pdf
DOI:
10.21014/tc10-2023.015
Event details
IMEKO TC:
TC10
Event name:
TC10 Conference 2023
Title:

19th IMEKO TC10 Conference "Measurement for Diagnostics, Optimisation and Control to Support Sustainability and Resilience"

Place:
Delft, The NETHERLANDS
Time:
21 September 2023 - 22 September 2023