Quantitative Detection of Metallic Particles in Transformer Oil Using Filter Paper Enrichment - LIBS

Online monitoring of metallic wear particles in transformer oil is of great significance for the early fault diagnosis of power equipment. However, conventional detection methods for trace particles suffer from uneven distribution and poor spectral stability. To address these issues, this study proposes a quantitative detection method for metallic particles in transformer oil based on laser-induced breakdown spectroscopy. An experimental platform was established, where vacuum filtration was employed to enrich Cu and Fe particles in oil. A baseline estimation algorithm based on spectral peak sparsity, combined with characteristic region selection and a Bagged-Trees classifier, was applied to achieve accurate baseline fitting, feature line extraction, and effective ablation point identification. Experimental results demonstrate that this method achieves a limit of quantification as low as 16 ng/g, with calibration curve correlation coefficients exceeding 0.995 and good repeatability. These findings verify that the proposed method enables highly sensitive and stable detection of trace metallic particles in transformer oil, providing a novel and effective approach for power equipment condition monitoring and fault prediction.