Open AccessIntelligent nonmodel-based fault diagnosis of electric motors using current signature analysis
Author(s) -
Ashraf A. Zaher,
Detlef Hummes,
G. Amjad Hussain
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1391/1/012065
Subject(s) - robustness (evolution) , computer science , harmonics , signature (topology) , induction motor , fault detection and isolation , fault (geology) , noise (video) , consistency (knowledge bases) , electric power system , control engineering , power (physics) , artificial intelligence , engineering , voltage , electrical engineering , mathematics , biochemistry , chemistry , geometry , physics , quantum mechanics , seismology , geology , actuator , image (mathematics) , gene
This paper proposes an efficient technique for detecting mechanical faults in three-phase induction motors, without using mechanical sensors. Only measurements of the currents of every phase are used to identify the fault. The proposed system can diagnose two types of faults corresponding to shaft misalignment or imbalance, along with normal operation. The power spectrum of the experimental data is generated, followed by applying a soft-computing mathematical algorithm that will extract the peaks of the fundamental frequencies and their harmonics, while filtering out noise. Carefully selected peaks at certain frequencies will be collected and examined to generate a robust algorithm that can be used to produce a decision regarding the operating condition of the motor, via applying an intelligent soft-computing technique. Mathematical details regarding the consistency checks for validating the experimental data, and the testing/validation phases will be investigated. Detailed analysis of the obtained results is provided to highlight the advantages and limitations of the proposed algorithm. In addition, a comparison is made with similar techniques that use mechanical sensors to contrast their differences and highlight the superiority of the proposed system. The obtained results prove the intelligence and robustness of the proposed system and allow for versatile extensions that promote its application in real-time scenarios for many industrial applications.