Diagnosing Electric Motor Faults based on Vibration Signals Using YOLOv8

The current demand in industries is increasingly focused on early diagnosis of faults in electric motors to prevent unwanted incidents during operation. Employing machine learning to monitor and detect potential faults helps mitigate catastrophic accidents and prevent losses of both assets and human...

Full description

Saved in:
Bibliographic Details
Published in2024 9th International Conference on Integrated Circuits, Design, and Verification (ICDV) pp. 191 - 196
Main Authors Tran, Hoai - Linh, Pham, Van - Nam, Nguyen, Duc-Thanh, Ba, Quang - Huy Do, Le, Xuan-Hai, Kim, Dinh-Thai
Format Conference Proceeding
LanguageEnglish
Published IEEE 06.06.2024
Subjects
Circuit faults
Faults in electric motors
Industries
Integrated circuit synthesis
Machine learning
Machine learning algorithms
Maintenance
Vibrations
YOLOv8
Online AccessGet full text

Cover

More Information
Summary:The current demand in industries is increasingly focused on early diagnosis of faults in electric motors to prevent unwanted incidents during operation. Employing machine learning to monitor and detect potential faults helps mitigate catastrophic accidents and prevent losses of both assets and human lives. A single fault occurrence can disrupt the motor's operation, leading to malfunctions in components such as coils, bearings, etc. Detecting faults using machine learning algorithms also serves as an effective method for preventive maintenance. This study utilizes the YOLOv8 algorithm to learn characteristics from the frequency distribution of vibration signals, targeting three common types of faults in electric motors during operation. By combining feature extraction and classification, it intelligently automates fault diagnosis. The paper also provides valuable insights into the performance of the YOLOv8 network and its practical applications in the field of maintenance and industrial equipment management.
DOI:10.1109/ICDV61346.2024.10617366