An Improved Fault Diagnosis Strategy for Induction Motors Using Weighted Probability Ensemble Deep Learning

Summary

The paper proposes a Weighted Probability Ensemble Deep Learning (WPEDL) approach for diagnosing faults in induction motors. WPEDL combines the strengths of multiple deep learning models to improve fault diagnosis accuracy.

Highlights

• WPEDL approach combines multiple deep learning models for improved fault diagnosis accuracy.
• The approach uses Short-Time Fourier Transform (STFT) to extract features from vibration and current signals.
• WPEDL achieves high accuracy in diagnosing bearing, rotor, and stator faults.
• The approach is compared with conventional deep learning models and shows superior performance.
• WPEDL is tested on a combined dataset of 52,000 STFT images and achieves an accuracy of 98.89%.
• The approach has potential applications in industrial settings for early fault detection and improved operational efficiency.
• WPEDL can handle high-dimensional data and provides a robust fault diagnosis solution.

Key Insights

• The WPEDL approach addresses the limitations of traditional fault diagnosis methods by combining the strengths of multiple deep learning models. This ensemble approach enables the model to learn complex patterns in the data and improve diagnosis accuracy.
• The use of STFT to extract features from vibration and current signals allows the model to capture time-frequency information, which is essential for diagnosing faults in induction motors.
• The WPEDL approach demonstrates high accuracy in diagnosing bearing, rotor, and stator faults, making it a robust solution for fault diagnosis in induction motors.
• The comparison with conventional deep learning models highlights the superiority of the WPEDL approach in terms of diagnosis accuracy.
• The WPEDL approach has significant potential for industrial applications, where early fault detection and diagnosis can improve operational efficiency and reduce maintenance costs.
• The ability of WPEDL to handle high-dimensional data makes it a suitable solution for complex fault diagnosis tasks.
• The WPEDL approach provides a robust fault diagnosis solution that can be used in various industrial settings, including power plants, manufacturing facilities, and transportation systems.

Mindmap

Citation

Ali, U., Ali, W., & Ramzan, U. (2024). An Improved Fault Diagnosis Strategy for Induction Motors Using Weighted Probability Ensemble Deep Learning (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2412.18249