Multi-Axis Accelerometer Based Evaluation and Classification of Misalignment, Soft Foot, and Looseness Faults in Induction Motors
IEEE Sensors Journal, vol.26, no.9, pp.14106-14116, 2026 (SCI-Expanded, Scopus)
- Publication Type: Article / Article
- Volume: 26 Issue: 9
- Publication Date: 2026
- Doi Number: 10.1109/jsen.2026.3678373
- Journal Name: IEEE Sensors Journal
- Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC
- Page Numbers: pp.14106-14116
- Keywords: Fault detection, induction motor (IM), looseness fault (Type B), parallel misalignment fault, soft foot fault, vibration analysis
- Dokuz Eylül University Affiliated: Yes
Abstract
In induction motors, possible faults can affect the dynamic performance of the motor and reduce the overall efficiency of the motor-driven systems. Therefore, it is crucial to detect potential faults early and accurately. In this study, parallel misalignment, soft foot, and looseness- Bolt (Type-B) faults have been investigated using multi-axis accelerometer data. For this purpose, a dedicated test setup is built to introduce different mechanical faults and analyze the frequency spectra of 3-axis vibration signals under varying torque profiles. Furthermore, a Random Forest–based artificial intelligence model has been developed to achieve highly accurate fault classification. Experimental results show that characteristic fault-related components in the vibration spectra such as fr (1X), 2fr (2X), 3fr (3X), and their sidebands (2fr+2sfs) and (3fr+2sfs) provide accurate and reliable fault detection. Moreover, the developed artificial intelligence model achieves fault classification with an accuracy ranging from 85.3% to 91.8%.