Moisture Prediction of Transformer Oil-Immersed Polymer Insulation by Applying a Support Vector Machine Combined with...

Area of Science:

• Electrical Engineering
• Materials Science
• Data Science

Background:

• Transformer condition assessment is vital for grid reliability.
• Moisture in oil-immersed insulation degrades performance.
• Existing methods for moisture prediction using GA-SVM are limited.

Purpose of the Study:

• To pioneer the application of GA-SVM and FDS for moisture prediction in transformer oil-immersed insulation.
• To develop a novel GA-SVM multi-classifier based on a fitting analysis model for moisture diagnosis.
• To validate the feasibility and reliability of the proposed method through experimental testing.

Main Methods:

• Utilizing the genetic algorithm-support vector machine (GA-SVM) for classification.
• Employing frequency domain spectroscopy (FDS) for data acquisition.
• Developing a fitting analysis model to construct a GA-SVM multi-classifier.
• Conducting laboratory and field tests for validation.

Main Results:

• Successfully applied GA-SVM and FDS for moisture prediction in transformer oil-immersed insulation.
• Demonstrated the feasibility and reliability of the developed GA-SVM multi-classifier.
• Experimental results confirm the model's potential for moisture diagnosis.

Conclusions:

• The GA-SVM combined with FDS presents a promising tool for moisture prediction in transformer oil-immersed polymer insulation.
• The developed multi-classifier model shows high accuracy and reliability.
• This method can enhance the condition assessment and maintenance strategies for transformers.