Electric Motor Vibration Signal Classification Using Wigner-Ville Distribution for Fault Diagnosis.

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Traditional fault diagnosis relies on time and frequency domain characteristics, which can be limited.
• Noise and vibration signal classification are crucial for mechanical and electronic systems, including electric vehicles.
• Accurate fault diagnosis in brushless motors is essential for reliability and stability.

Purpose of the Study:

• To propose and validate a new technique for visualizing and classifying vibration signals from brushless motors.
• To extract vibration signal characteristics using the Wigner-Ville Distribution (WVD) method.
• To utilize artificial neural networks, specifically the YOLO (you only look once) deep learning model, for signal classification.

Main Methods:

• Vibration signals were measured from a brushless motor operating under six different revolution states.
• The Wigner-Ville Distribution (WVD) method was employed to convert and visualize the vibration signals into images.
• The YOLO (you only look once) deep learning model was used to identify and classify these WVD images for fault diagnosis.

Main Results:

• The WVD method effectively extracted distinct vibration signal characteristics from the brushless motor.
• The YOLO deep learning model successfully identified and classified the WVD images, demonstrating the feasibility of the approach.
• Analysis of Wagener method parameters and recognition rates indicated improved motor fault diagnostic capabilities.

Conclusions:

• The proposed method enables accurate fault diagnosis of brushless motors without the need for dismantling.
• This technique enhances the reliability and stability of brushless motor applications through improved diagnostic accuracy.
• The integration of WVD visualization and YOLO deep learning offers a promising advancement in condition monitoring and fault diagnosis.