Study on the aging status of insulators based on hyperspectral imaging technology

Area of Science:

• Materials Science
• Electrical Engineering
• Remote Sensing

Background:

• Acidic environments degrade silicone rubber (SiR) insulators, reducing hydrophobicity and flashover resistance, posing risks to power grid stability.
• Accurate assessment of insulator aging is crucial for preventing transmission line failures.

Purpose of the Study:

• To develop a pixel-level assessment method for SiR insulator aging using optical hyperspectral imaging (HSI).
• To evaluate the effectiveness of a combined RF-PCA and CNN approach for analyzing HSI data of aged insulators.

Main Methods:

• Artificial aging of SiR samples in acidic solutions (HNO3, H2SO4, HCl) to create six aging grades.
• Acquisition of hyperspectral images (HSI) for each aging grade.
• Application of a joint random forest-principal component analysis (RF-PCA) for dimensionality reduction from 256 to 7 dimensions.
• Development of a convolutional neural network (CNN) model for pixel-level classification and defect prediction.

Main Results:

• The RF-PCA method effectively reduced data complexity while retaining essential spectral information.
• The CNN model achieved accurate pixel-level classification of SiR insulator aging states.
• The methodology demonstrated potential for visual prediction of insulator defects.

Conclusions:

• HSI combined with RF-PCA and CNN offers a robust method for assessing SiR insulator aging.
• This approach enhances the timely detection of power grid safety hazards.
• The study provides a foundation for proactive maintenance and improved power grid reliability.