Research on a Transformer Vibration Fault Diagnosis Method Based on Time-Shift Multiscale Increment Entropy and...

Area of Science:

• Electrical Engineering
• Signal Processing
• Machine Learning

Background:

• Transformer mechanical vibration analysis is crucial for safe operation.
• Existing multiscale entropy methods can lose information during coarse-graining.
• Accurate fault diagnosis is essential for preventing transformer failures.

Purpose of the Study:

• To propose a novel fault diagnosis method for transformer mechanical vibrations.
• To improve the accuracy and stability of transformer fault diagnosis.
• To introduce time-shift multiscale increment entropy (TSMIE) as an effective feature extraction technique.

Main Methods:

• Developed time-shift multiscale increment entropy (TSMIE) to overcome information loss in traditional methods.
• Extracted TSMIE features from transformer vibration signals under various operating conditions.
• Utilized the CatBoost model for pattern recognition and fault classification.

Main Results:

• The proposed TSMIE combined with CatBoost demonstrated higher diagnostic accuracy and stability.
• Simulation and experimental results validated the effectiveness of the new method.
• The method offers a significant improvement over existing transformer vibration fault diagnosis techniques.

Conclusions:

• The novel TSMIE feature extraction combined with CatBoost provides an effective tool for transformer vibration fault diagnosis.
• This approach enhances diagnostic accuracy and operational stability.
• It represents a valuable advancement in ensuring the reliable operation of power transformers.