An Evaluation Method of Composite Insulator Aging Grade Based on Probabilistic Neural Network

This paper proposes a hyperspectral-based method for evaluating the corona aging state of composite insulators used in the distribution of electricity. The traditional testing method requires offline conditions and results in a significant economic burden. The proposed method involves processing hyperspectral data using PCA for dimensionality reduction and then classifying composite insulator samples with different aging states using probabilistic neural networks. This approach enables real-time monitoring of the aging state of composite insulators, reducing the risk of equipment failure and enhancing the reliability of power transmission. Moreover, this approach can help reduce the economic burden associated with traditional testing methods. The study highlights the potential of hyperspectral-based methods for evaluating the corona aging state of composite insulators, with significant implications for the power industry.